Researcher profile

Weinan Zhang

Weinan Zhang contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate

Artificial Intelligence Machine Learning Information Retrieval Computation and Language Multiagent Systems cs.CY Software Engineering Computer Vision Computer Science and Game Theory Digital Libraries Discrete Mathematics math.OC Programming Languages Robotics Social and Information Networks

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author

60works

0followers

15topics

4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Adding Alignment Control to Language Models

Post-training alignment has increasingly become a crucial factor in enhancing the usability of language models (LMs). However, the strength of alignment varies depending on individual preferences. This paper proposes a method to incorporate alignment control into a single model, referred to as CLM. This approach adds one identity layer preceding the initial layers and performs preference learning only on this layer to map unaligned input token embeddings into the aligned space. Experimental results demonstrate that this efficient fine-tuning method performs comparable to full fine-tuning. During inference, the input embeddings are processed through the aligned and unaligned layers, which are then merged through the interpolation coefficient. By controlling this parameter, the alignment exhibits a clear interpolation and extrapolation phenomenon.

preprint2026arXiv

AdvKT: An Adversarial Multi-Step Training Framework for Knowledge Tracing

Knowledge Tracing (KT) monitors students' knowledge states and simulates their responses to question sequences. Existing KT models typically follow a single-step training paradigm, which leads to discrepancies with the multi-step inference process required in real-world simulations, resulting in significant error accumulation. This accumulation of error, coupled with the issue of data sparsity, can substantially degrade the performance of recommendation models in the intelligent tutoring systems. To address these challenges, we propose a novel Adversarial Multi-Step Training Framework for Knowledge Tracing (AdvKT), which, for the first time, focuses on the multi-step KT task. More specifically, AdvKT leverages adversarial learning paradigm involving a generator and a discriminator. The generator mimics high-reward responses, effectively reducing error accumulation across multiple steps, while the discriminator provides feedback to generate synthetic data. Additionally, we design specialized data augmentation techniques to enrich the training data with realistic variations, ensuring that the model generalizes well even in scenarios with sparse data. Experiments conducted on four real-world datasets demonstrate the superiority of AdvKT over existing KT models, showcasing its ability to address both error accumulation and data sparsity issues effectively.

preprint2026arXiv

Contexting as Recommendation: Evolutionary Collaborative Filtering for Context Engineering

Large Language Models (LLMs) are highly sensitive to their input contexts, motivating the development of automated context engineering. However, existing methods predominantly treat this as a global search problem, seeking a single context strategy that maximizes average performance across a dataset. This restrictive assumption overlooks the fact that different inputs often require distinct guidance, leaving substantial instance-level performance gains untapped. In this paper, we propose a paradigm shift by formulating context engineering as a recommendation problem. We introduce \textbf{Neural Collaborative Context Engineering (NCCE)}, a framework that transitions optimization from a static global search to dynamic, instance-wise routing. NCCE first bootstraps a diverse catalog of anchor contexts and then employs a novel \textbf{Context-CF Co-Evolution} mechanism. This stage establishes a synergistic feedback loop: a lightweight Neural Collaborative Filtering (NCF) model learns instance-context preferences to guide the generation of specialized context variants, while the newly evaluated contexts continuously refine the NCF model's understanding of latent preferences. At inference time, the trained NCF model acts as a context router, dynamically assigning the most suitable context strategy to each unseen instance. Theoretical Proofs and comprehensive experiments demonstrate that by matching individual inputs with their optimal contexts, NCCE significantly improves task accuracy, highlighting the critical importance of personalization in LLM context engineering.

preprint2026arXiv

CoreCodeBench: Decoupling Code Intelligence via Fine-Grained Repository-Level Tasks

The evaluation of Large Language Models (LLMs) for software engineering has shifted towards complex, repository-level tasks. However, existing benchmarks predominantly rely on coarse-grained pass rates that treat programming proficiency as a monolithic capability, obscuring specific cognitive bottlenecks. Furthermore, the static nature of these benchmarks renders them vulnerable to data contamination and performance saturation. To address these limitations, we introduce CoreCodeBench, a configurable repository-level benchmark designed to dissect coding capabilities through atomized tasks. Leveraging our automated framework, CorePipe, we extract and transform Python repositories into a comprehensive suite of tasks that isolate distinct cognitive demands within identical code contexts. Unlike static evaluations, CoreCodeBench supports controllable difficulty scaling to prevent saturation and ensures superior data quality. It achieves a 78.55% validity yield, significantly surpassing the 31.7% retention rate of SWE-bench-Verified. Extensive experiments with state-of-the-art LLMs reveal a significant capability misalignment, evidenced by distinct ranking shifts across cognitive dimensions. This indicates that coding proficiency is non-monolithic, as strength in one aspect does not necessarily translate to others. These findings underscore the necessity of our fine-grained taxonomy in diagnosing model deficiencies and offer a sustainable, rigorous framework for evolving code intelligence. The code for CorePipe is available at https://github.com/AGI-Eval-Official/CoreCodeBench, and the data for CoreCodeBench can be accessed at https://huggingface.co/collections/tubehhh/corecodebench-68256d2faabf4b1610a08caa.

preprint2026arXiv

Flexible Realignment of Language Models

Realignment becomes necessary when a language model (LM) fails to meet expected performance. We propose a flexible realignment framework that supports quantitative control of alignment degree during training and inference. This framework incorporates Training-time Realignment (TrRa), which efficiently realigns the reference model by leveraging the controllable fusion of logits from both the reference and already aligned models. For example, TrRa reduces token usage by 54.63% on DeepSeek-R1-Distill-Qwen-1.5B without any performance degradation, outperforming DeepScaleR-1.5B's 33.86%. To complement TrRa during inference, we introduce a layer adapter that enables smooth Inference-time Realignment (InRa). This adapter is initialized to perform an identity transformation at the bottom layer and is inserted preceding the original layers. During inference, input embeddings are simultaneously processed by the adapter and the original layer, followed by the remaining layers, and then controllably interpolated at the logit level. We upgraded DeepSeek-R1-Distill-Qwen-7B from a slow-thinking model to one that supports both fast and slow thinking, allowing flexible alignment control even during inference. By encouraging deeper reasoning, it even surpassed its original performance.

preprint2026arXiv

Gradients with Respect to Semantics Preserving Embeddings Tell the Uncertainty of Large Language Models

Uncertainty quantification (UQ) is an important technique for ensuring the trustworthiness of LLMs, given their tendency to hallucinate. Existing state-of-the-art UQ approaches for free-form generation rely heavily on sampling, which incurs high computational cost and variance. In this work, we propose the first gradient-based UQ method for free-form generation, SemGrad, which is sampling-free and computationally efficient. Unlike prior gradient-based methods developed for classification tasks that operates in parameter space, we propose to consider gradients in semantic space. Our method builds on the key intuition that a confident LLM should maintain stable output distributions under semantically equivalent input perturbations. We interpret the stability as the gradients in semantic space and introduce a Semantic Preservation Score (SPS) to identify embeddings that best capture semantics, with respect to which gradients are computed. We further propose HybridGrad, which combines the strengths of SemGrad and parameter gradients. Experiments demonstrate that both of our methods provide efficient and effective uncertainty estimates, achieving superior performance than state-of-the-art methods, particularly in settings with multiple valid responses.

preprint2026arXiv

Hölder Policy Optimisation

Group Relative Policy Optimisation (GRPO) enhances large language models by estimating advantages across a group of sampled trajectories. However, mapping these trajectory-level advantages to policy updates requires aggregating token-level probabilities within each sequence. Relying on a fixed aggregation mechanism for this step fundamentally limits the algorithm's adaptability. Empirically, we observe a critical trade-off: certain fixed aggregations frequently suffer from training collapse, while others fail to yield satisfactory performance. To resolve this, we propose \textbf{HölderPO}, a generalised policy optimisation framework unifying token-level probability aggregation via the Hölder mean. By explicitly modulating the parameter $p$, our framework provides continuous control over the trade-off between gradient concentration and variance bounds. Theoretically, we prove that a larger $p$ concentrates the gradient to amplify sparse learning signals, whereas a smaller $p$ strictly bounds gradient variance. Because no static configuration can universally resolve this concentration-stability trade-off, we instantiate the framework with a dynamic annealing algorithm that progressively schedules $p$ across the training lifecycle. Extensive evaluations demonstrate superior stability and convergence over existing baselines. Specifically, our approach achieves a state-of-the-art average accuracy of $54.9\%$ across multiple mathematical benchmarks, yielding a substantial $7.2\%$ relative gain over standard GRPO and secures an exceptional $93.8\%$ success rate on ALFWorld.

preprint2026arXiv

MemQ: Integrating Q-Learning into Self-Evolving Memory Agents over Provenance DAGs

Episodic memory allows LLM agents to accumulate and retrieve experience, but current methods treat each memory independently, i.e., evaluating retrieval quality in isolation without accounting for the dependency chains through which memories enable the creation of future memories. We introduce MemQ, which applies TD($λ$) eligibility traces to memory Q-values, propagating credit backward through a provenance DAG that records which memories were retrieved when each new memory was created. Credit weight decays as $(γλ)^d$ with DAG depth $d$, replacing temporal distance with structural proximity. We formalize the setting as an Exogenous-Context MDP, whose factored transition decouples the exogenous task stream from the endogenous memory store. Across six benchmarks, spanning OS interaction, function calling, code generation, multimodal reasoning, embodied reasoning, and expert-level QA, MemQ achieves the highest success rate on all six in generalization evaluation and runtime learning, with gains largest on multi-step tasks that produce deep and relevant provenance chains (up to +5.7~pp) and smallest on single-step classification (+0.77~pp) where single-step updates already suffice. We further study how $γ$ and $λ$ interact with the EC-MDP structure, providing principled guidance for parameter selection and future research. Code is available at https://github.com/jwliao-ai/MemQ.

preprint2026arXiv

MMSkills: Towards Multimodal Skills for General Visual Agents

Reusable skills have become a core substrate for improving agent capabilities, yet most existing skill packages encode reusable behavior primarily as textual prompts, executable code, or learned routines. For visual agents, however, procedural knowledge is inherently multimodal: reuse depends not only on what operation to perform, but also on recognizing the relevant state, interpreting visual evidence of progress or failure, and deciding what to do next. We formalize this requirement as multimodal procedural knowledge and address three practical challenges: (I) what a multimodal skill package should contain; (II) where such packages can be derived from public interaction experience; and (III) how agents can consult multimodal evidence at inference time without excessive image context or over-anchoring to reference screenshots. We introduce MMSkills, a framework for representing, generating, and using reusable multimodal procedures for runtime visual decision making. Each MMSkill is a compact, state-conditioned package that couples a textual procedure with runtime state cards and multi-view keyframes. To construct these packages, we develop an agentic trajectory-to-skill Generator that transforms public non-evaluation trajectories into reusable multimodal skills through workflow grouping, procedure induction, visual grounding, and meta-skill-guided auditing. To use them, we introduce a branch-loaded multimodal skill agent: selected state cards and keyframes are inspected in a temporary branch, aligned with the live environment, and distilled into structured guidance for the main agent. Experiments across GUI and game-based visual-agent benchmarks show that MMSkills consistently improve both frontier and smaller multimodal agents, suggesting that external multimodal procedural knowledge complements model-internal priors.

preprint2026arXiv

Position: Academic Conferences are Potentially Facing Denominator Gaming Caused by Fully Automated Scientific Agents

The implicit policy of maintaining relatively stable acceptance rates at top AI conferences, despite exponentially growing submissions, introduces a critical structural vulnerability. This position paper characterizes a new systemic threat we term Agentic Denominator Gaming, in which a malicious actor deploys AI agents to generate and submit a large volume of superficially plausible but low-quality papers. Crucially, their objective is not the acceptance of low-quality papers, but rather to inflate the submission denominator and overwhelm reviewing capacity. Under a relatively stable acceptance rate, this dilution can systematically increase the publication probability of a small, targeted set of legitimate papers. We analyze the practical feasibility of this threat and its broader consequences, including intensified reviewer burnout, degraded review quality, and the emergence of industrialized automated agent mills. Finally, we propose and evaluate a range of mitigation strategies, and argue that durable protection will require system-level policy and incentive reforms, rather than relying primarily on technical detection alone.

preprint2026arXiv

Position: Agentic AI System Is a Foreseeable Pathway to AGI

Is monolithic scaling the only path to AGI? This paper challenges the dogma that purely scaling a single model is sufficient to achieve Artificial General Intelligence. Instead, we identify Agentic AI as a necessary paradigm for mastering the complex, heterogeneous distribution of real-world tasks. Through rigorous theoretical derivations, we contrast the optimization constraints of monolithic learners against the efficiency of Agentic systems, progressing from simple routing mechanisms to general Directed Acyclic Graph (DAG) topologies. We demonstrate that Agentic AI achieves exponentially superior generalization and sample efficiency. Finally, we discuss the connection to Mixture-of-Experts, reinterpret the instability of current multi-agent frameworks, and call for greater research focus on Agentic AI.

preprint2026arXiv

PRTS: A Primitive Reasoning and Tasking System via Contrastive Representations

Vision-Language-Action (VLA) models advance robotic control via strong visual-linguistic priors. However, existing VLAs predominantly frame pretraining as supervised behavior cloning, overlooking the fundamental nature of robot learning as a goal-reaching process that requires understanding temporal task progress. We present \textbf{PRTS} (\textbf{P}rimitive \textbf{R}easoning and \textbf{T}asking \textbf{S}ystem), a VLA foundation model that reformulates pretraining through Goal-Conditioned Reinforcement Learning. By treating language instructions as goals and employing contrastive reinforcement learning, PRTS learns a unified embedding space where the inner product of state-action and goal embeddings approximates the log-discounted goal occupancy, the probability of reaching the language-specified goal from the current state-action, quantitatively assessing physical feasibility beyond static semantic matching. PRTS draws this dense goal-reachability supervision directly from offline trajectories without reward annotations, and folds it into the VLM backbone via a role-aware causal mask, incurring negligible overhead over vanilla behavior cloning. This paradigm endows the high-level reasoning system with intrinsic goal reachability awareness, bridging semantic reasoning and temporal task progress, and further benefits goal-conditioned action prediction. Pretrained on 167B tokens of diverse manipulation and embodied-reasoning data, PRTS reaches state-of-the-art performance on LIBERO, LIBERO-Pro, LIBERO-Plus, SimplerEnv, and a real-world suite of 14 complex tasks, with particularly substantial gains on long-horizon, contact-rich, and zero-shot novel-instruction settings, confirming that injecting goal-reachability awareness significantly improves both execution success and long-horizon planning of general-purpose robotic foundation policies.

preprint2026arXiv

SkillMAS: Skill Co-Evolution with LLM-based Multi-Agent System

Large language model (LLM) agent systems are increasingly expected to improve after deployment, but existing work often decouples two adaptation targets: skill evolution and multi-agent system (MAS) restructuring. This separation can create organization bottlenecks, context pressure, and mis-specialization. We present SkillMAS, a non-parametric framework for adaptive specialization in multi-agent systems that couples skill evolution with MAS restructuring. SkillMAS uses Utility Learning to assign credit from verified execution traces, bounded skill evolution to refine reusable procedures without unfiltered library growth, and evidence-gated MAS restructuring when retained failures and Executor Utility indicate a structural mismatch. Across embodied manipulation, command-line execution, and retail workflows, SkillMAS is competitive under the reported harnesses while clarifying how post-deployment specialization is attributed, updated, and applied.

preprint2026arXiv

Skills on the Fly: Test-Time Adaptive Skill Synthesis for LLM Agents

LLM agents benefit from reusable skills, yet test-time tasks often require guidance more specific than a static skill library can provide. We propose \emph{SkillTTA}, a Test-Time Adaptive Skill Synthesis method that retrieves a small set of training trajectories relevant to the current task and synthesizes them into a temporary, task-specific textual skill. The solver model is kept fixed, so adaptation happens entirely through generated context rather than parameter updates. We evaluate the method on SpreadsheetBench, ALFWorld, and BigCodeBench. Compared with static trajectory-to-skill synthesis using GPT-5.5, task-specific skills improve SpreadsheetBench Pass@1 from 0.397 to 0.505 and BigCodeBench Pass@1 from 0.517 to 0.651. On ALFWorld, the method matches a heavier memory-learning baseline within four points of success rate while producing the shortest successful trajectories among reported methods. Ablations on SpreadsheetBench further show that synthesized skills outperform raw trajectory prompting, that top-$k$ retrieval should stay small, and that failed trajectories are especially useful because they expose recurring evaluator-facing mistakes.

preprint2026arXiv

SMMBench: A Benchmark for Source-Distributed Multimodal Agent Memory

Existing benchmarks for multimodal memory reasoning largely evaluate systems within pre-assembled contexts, but under-evaluate whether agents can use evidence distributed across independently originated sources. We argue that source-distributed memory composition is an important and under-examined bottleneck in multimodal agent memory, especially when relevant evidence is fragmented across heterogeneous artifacts such as conversations, profiles, screenshots, tables, images, and documents. To address this gap, we introduce Source-distributed Multimodal Memory Benchmark(SMMBench), which measures whether agents can retrieve, align, and compose multimodal evidence scattered across multiple sources rather than reason within a single curated context. SMMBench evaluates four core capabilities: (1) cross-source multimodal reasoning; (2) conflict resolution; (3) preference reasoning; (4) memory-grounded action prediction. The benchmark contains 1877 samples grounded in 264 sources. Experiments on representative memory-style and retrieval-based baselines show that current systems still struggle on these capabilities, positioning source-distributed multimodal memory as an important and still under-evaluated challenge for multimodal agents. Our data are available at https://huggingface.co/datasets/HuacanChai/SMMBench.

preprint2026arXiv

XSearch: Explainable Code Search via Concept-to-Code Alignment

Semantic code search has been widely adopted in both academia and industry. These approaches embed natural-language queries and code snippets into a shared embedding space and retrieve results based on vector similarity. Despit strong performance on benchmark datasets, they often suffer from poor explainability and generalization. Retrieved code may appear semantically similar yet miss critical functional requirements of the query, while providing no explanation of why the result was retrieved. Moreover, such failures become more severe under distribution shift, where models struggle to generalize to unseen benchmarks. In this work, we propose XSearch, an intrinsically explainable code search framework. Our key insight is that by relying on global embedding similarity, existing retrievers inherently take an inductive view. They learn statistical patterns rather than truly understanding the query's functional requirements. We address this problem by reformulating code search as a deductive concept alignment problem. XSearch (i) identifies functional concepts in the query and (ii) explicitly aligns them with corresponding code statements. This explain-then-predict design produces inherent concept-level explanations and mitigates shortcut learning that harms out-of-distribution generalization. We train an encoder with explicit concept-alignment objectives and perform retrieval through explicit matching between query concepts and code statements. Experiments show that, trained on CodeSearchNet using GraphCodeBERT (125M parameters), XSearch improves performance on out-of-distribution benchmarks from 0.02 to 0.33 (15x) over eight state-of-the-art retrievers, and consistently outperforms both encoder- and decoder-based baselines with up to 7B parameters. A user study demonstrates that concept-alignment explanations enable users to evaluate retrieved results faster and more accurately.

preprint2022arXiv

A Graph-Enhanced Click Model for Web Search

To better exploit search logs and model users' behavior patterns, numerous click models are proposed to extract users' implicit interaction feedback. Most traditional click models are based on the probabilistic graphical model (PGM) framework, which requires manually designed dependencies and may oversimplify user behaviors. Recently, methods based on neural networks are proposed to improve the prediction accuracy of user behaviors by enhancing the expressive ability and allowing flexible dependencies. However, they still suffer from the data sparsity and cold-start problems. In this paper, we propose a novel graph-enhanced click model (GraphCM) for web search. Firstly, we regard each query or document as a vertex, and propose novel homogeneous graph construction methods for queries and documents respectively, to fully exploit both intra-session and inter-session information for the sparsity and cold-start problems. Secondly, following the examination hypothesis, we separately model the attractiveness estimator and examination predictor to output the attractiveness scores and examination probabilities, where graph neural networks and neighbor interaction techniques are applied to extract the auxiliary information encoded in the pre-constructed homogeneous graphs. Finally, we apply combination functions to integrate examination probabilities and attractiveness scores into click predictions. Extensive experiments conducted on three real-world session datasets show that GraphCM not only outperforms the state-of-art models, but also achieves superior performance in addressing the data sparsity and cold-start problems.

preprint2022arXiv

A Roadmap for Big Model

With the rapid development of deep learning, training Big Models (BMs) for multiple downstream tasks becomes a popular paradigm. Researchers have achieved various outcomes in the construction of BMs and the BM application in many fields. At present, there is a lack of research work that sorts out the overall progress of BMs and guides the follow-up research. In this paper, we cover not only the BM technologies themselves but also the prerequisites for BM training and applications with BMs, dividing the BM review into four parts: Resource, Models, Key Technologies and Application. We introduce 16 specific BM-related topics in those four parts, they are Data, Knowledge, Computing System, Parallel Training System, Language Model, Vision Model, Multi-modal Model, Theory&Interpretability, Commonsense Reasoning, Reliability&Security, Governance, Evaluation, Machine Translation, Text Generation, Dialogue and Protein Research. In each topic, we summarize clearly the current studies and propose some future research directions. At the end of this paper, we conclude the further development of BMs in a more general view.

preprint2022arXiv

A Survey on Model-based Reinforcement Learning

Reinforcement learning (RL) solves sequential decision-making problems via a trial-and-error process interacting with the environment. While RL achieves outstanding success in playing complex video games that allow huge trial-and-error, making errors is always undesired in the real world. To improve the sample efficiency and thus reduce the errors, model-based reinforcement learning (MBRL) is believed to be a promising direction, which builds environment models in which the trial-and-errors can take place without real costs. In this survey, we take a review of MBRL with a focus on the recent progress in deep RL. For non-tabular environments, there is always a generalization error between the learned environment model and the real environment. As such, it is of great importance to analyze the discrepancy between policy training in the environment model and that in the real environment, which in turn guides the algorithm design for better model learning, model usage, and policy training. Besides, we also discuss the recent advances of model-based techniques in other forms of RL, including offline RL, goal-conditioned RL, multi-agent RL, and meta-RL. Moreover, we discuss the applicability and advantages of MBRL in real-world tasks. Finally, we end this survey by discussing the promising prospects for the future development of MBRL. We think that MBRL has great potential and advantages in real-world applications that were overlooked, and we hope this survey could attract more research on MBRL.

preprint2022arXiv

Branch Ranking for Efficient Mixed-Integer Programming via Offline Ranking-based Policy Learning

Deriving a good variable selection strategy in branch-and-bound is essential for the efficiency of modern mixed-integer programming (MIP) solvers. With MIP branching data collected during the previous solution process, learning to branch methods have recently become superior over heuristics. As branch-and-bound is naturally a sequential decision making task, one should learn to optimize the utility of the whole MIP solving process instead of being myopic on each step. In this work, we formulate learning to branch as an offline reinforcement learning (RL) problem, and propose a long-sighted hybrid search scheme to construct the offline MIP dataset, which values the long-term utilities of branching decisions. During the policy training phase, we deploy a ranking-based reward assignment scheme to distinguish the promising samples from the long-term or short-term view, and train the branching model named Branch Ranking via offline policy learning. Experiments on synthetic MIP benchmarks and real-world tasks demonstrate that Branch Rankink is more efficient and robust, and can better generalize to large scales of MIP instances compared to the widely used heuristics and state-of-the-art learning-based branching models.

preprint2022arXiv

Context-aware Reranking with Utility Maximization for Recommendation

As a critical task for large-scale commercial recommender systems, reranking has shown the potential of improving recommendation results by uncovering mutual influence among items. Reranking rearranges items in the initial ranking lists from the previous ranking stage to better meet users' demands. However, rather than considering the context of initial lists as most existing methods do, an ideal reranking algorithm should consider the counterfactual context -- the position and the alignment of the items in the reranked lists. In this work, we propose a novel pairwise reranking framework, Context-aware Reranking with Utility Maximization for recommendation (CRUM), which maximizes the overall utility after reranking efficiently. Specifically, we first design a utility-oriented evaluator, which applies Bi-LSTM and graph attention mechanism to estimate the listwise utility via the counterfactual context modeling. Then, under the guidance of the evaluator, we propose a pairwise reranker model to find the most suitable position for each item by swapping misplaced item pairs. Extensive experiments on two benchmark datasets and a proprietary real-world dataset demonstrate that CRUM significantly outperforms the state-of-the-art models in terms of both relevance-based metrics and utility-based metrics.

preprint2022arXiv

Curriculum Offline Imitation Learning

Offline reinforcement learning (RL) tasks require the agent to learn from a pre-collected dataset with no further interactions with the environment. Despite the potential to surpass the behavioral policies, RL-based methods are generally impractical due to the training instability and bootstrapping the extrapolation errors, which always require careful hyperparameter tuning via online evaluation. In contrast, offline imitation learning (IL) has no such issues since it learns the policy directly without estimating the value function by bootstrapping. However, IL is usually limited in the capability of the behavioral policy and tends to learn a mediocre behavior from the dataset collected by the mixture of policies. In this paper, we aim to take advantage of IL but mitigate such a drawback. Observing that behavior cloning is able to imitate neighboring policies with less data, we propose \textit{Curriculum Offline Imitation Learning (COIL)}, which utilizes an experience picking strategy for imitating from adaptive neighboring policies with a higher return, and improves the current policy along curriculum stages. On continuous control benchmarks, we compare COIL against both imitation-based and RL-based methods, showing that it not only avoids just learning a mediocre behavior on mixed datasets but is also even competitive with state-of-the-art offline RL methods.

preprint2022arXiv

DropNAS: Grouped Operation Dropout for Differentiable Architecture Search

Neural architecture search (NAS) has shown encouraging results in automating the architecture design. Recently, DARTS relaxes the search process with a differentiable formulation that leverages weight-sharing and SGD where all candidate operations are trained simultaneously. Our empirical results show that such procedure results in the co-adaption problem and Matthew Effect: operations with fewer parameters would be trained maturely earlier. This causes two problems: firstly, the operations with more parameters may never have the chance to express the desired function since those with less have already done the job; secondly, the system will punish those underperforming operations by lowering their architecture parameter, and they will get smaller loss gradients, which causes the Matthew Effect. In this paper, we systematically study these problems and propose a novel grouped operation dropout algorithm named DropNAS to fix the problems with DARTS. Extensive experiments demonstrate that DropNAS solves the above issues and achieves promising performance. Specifically, DropNAS achieves 2.26% test error on CIFAR-10, 16.39% on CIFAR-100 and 23.4% on ImageNet (with the same training hyperparameters as DARTS for a fair comparison). It is also observed that DropNAS is robust across variants of the DARTS search space. Code is available at https://github.com/wiljohnhong/DropNAS.

preprint2022arXiv

Efficient Policy Space Response Oracles

Policy Space Response Oracle methods (PSRO) provide a general solution to learn Nash equilibrium in two-player zero-sum games but suffer from two drawbacks: (1) the computation inefficiency due to the need for consistent meta-game evaluation via simulations, and (2) the exploration inefficiency due to finding the best response against a fixed meta-strategy at every epoch. In this work, we propose Efficient PSRO (EPSRO) that largely improves the efficiency of the above two steps. Central to our development is the newly-introduced subroutine of no-regret optimization on the unrestricted-restricted (URR) game. By solving URR at each epoch, one can evaluate the current game and compute the best response in one forward pass without the need for meta-game simulations. Theoretically, we prove that the solution procedures of EPSRO offer a monotonic improvement on the exploitability, which none of existing PSRO methods possess. Furthermore, we prove that the no-regret optimization has a regret bound of $\mathcal{O}(\sqrt{T\log{[(k^2+k)/2]}})$, where $k$ is the size of restricted policy set. Most importantly, a desirable property of EPSRO is that it is parallelizable, this allows for highly efficient exploration in the policy space that induces behavioral diversity. We test EPSRO on three classes of games, and report a 50x speedup in wall-time and 10x data efficiency while maintaining similar exploitability as existing PSRO methods on Kuhn and Leduc Poker games.

preprint2022arXiv

Forgetting Fast in Recommender Systems

Users of a recommender system may want part of their data being deleted, not only from the data repository but also from the underlying machine learning model, for privacy or utility reasons. Such right-to-be-forgotten requests could be fulfilled by simply retraining the recommendation model from scratch, but that would be too slow and too expensive in practice. In this paper, we investigate fast machine unlearning techniques for recommender systems that can remove the effect of a small amount of training data from the recommendation model without incurring the full cost of retraining. A natural idea to speed this process up is to fine-tune the current recommendation model on the remaining training data instead of starting from a random initialization. This warm-start strategy indeed works for neural recommendation models using standard 1st-order neural network optimizers (like AdamW). However, we have found that even greater acceleration could be achieved by employing 2nd-order (Newton or quasi-Newton) optimization methods instead. To overcome the prohibitively high computational cost of 2nd-order optimizers, we propose a new recommendation unlearning approach AltEraser which divides the optimization problem of unlearning into many small tractable sub-problems. Extensive experiments on three real-world recommendation datasets show promising results of AltEraser in terms of consistency (forgetting thoroughness), accuracy (recommendation effectiveness), and efficiency (unlearning speed). To our knowledge, this work represents the first attempt at fast approximate machine unlearning for state-of-the-art neural recommendation models.

preprint2022arXiv

Generative Adversarial Exploration for Reinforcement Learning

Exploration is crucial for training the optimal reinforcement learning (RL) policy, where the key is to discriminate whether a state visiting is novel. Most previous work focuses on designing heuristic rules or distance metrics to check whether a state is novel without considering such a discrimination process that can be learned. In this paper, we propose a novel method called generative adversarial exploration (GAEX) to encourage exploration in RL via introducing an intrinsic reward output from a generative adversarial network, where the generator provides fake samples of states that help discriminator identify those less frequently visited states. Thus the agent is encouraged to visit those states which the discriminator is less confident to judge as visited. GAEX is easy to implement and of high training efficiency. In our experiments, we apply GAEX into DQN and the DQN-GAEX algorithm achieves convincing performance on challenging exploration problems, including the game Venture, Montezuma's Revenge and Super Mario Bros, without further fine-tuning on complicate learning algorithms. To our knowledge, this is the first work to employ GAN in RL exploration problems.

preprint2022arXiv

Geometer: Graph Few-Shot Class-Incremental Learning via Prototype Representation

With the tremendous expansion of graphs data, node classification shows its great importance in many real-world applications. Existing graph neural network based methods mainly focus on classifying unlabeled nodes within fixed classes with abundant labeling. However, in many practical scenarios, graph evolves with emergence of new nodes and edges. Novel classes appear incrementally along with few labeling due to its newly emergence or lack of exploration. In this paper, we focus on this challenging but practical graph few-shot class-incremental learning (GFSCIL) problem and propose a novel method called Geometer. Instead of replacing and retraining the fully connected neural network classifer, Geometer predicts the label of a node by finding the nearest class prototype. Prototype is a vector representing a class in the metric space. With the pop-up of novel classes, Geometer learns and adjusts the attention-based prototypes by observing the geometric proximity, uniformity and separability. Teacher-student knowledge distillation and biased sampling are further introduced to mitigate catastrophic forgetting and unbalanced labeling problem respectively. Experimental results on four public datasets demonstrate that Geometer achieves a substantial improvement of 9.46% to 27.60% over state-of-the-art methods.

preprint2022arXiv

Goal-Conditioned Reinforcement Learning: Problems and Solutions

Goal-conditioned reinforcement learning (GCRL), related to a set of complex RL problems, trains an agent to achieve different goals under particular scenarios. Compared to the standard RL solutions that learn a policy solely depending on the states or observations, GCRL additionally requires the agent to make decisions according to different goals. In this survey, we provide a comprehensive overview of the challenges and algorithms for GCRL. Firstly, we answer what the basic problems are studied in this field. Then, we explain how goals are represented and present how existing solutions are designed from different points of view. Finally, we make the conclusion and discuss potential future prospects that recent researches focus on.

preprint2022arXiv

Learn over Past, Evolve for Future: Search-based Time-aware Recommendation with Sequential Behavior Data

The personalized recommendation is an essential part of modern e-commerce, where user's demands are not only conditioned by their profile but also by their recent browsing behaviors as well as periodical purchases made some time ago. In this paper, we propose a novel framework named Search-based Time-Aware Recommendation (STARec), which captures the evolving demands of users over time through a unified search-based time-aware model. More concretely, we first design a search-based module to retrieve a user's relevant historical behaviors, which are then mixed up with her recent records to be fed into a time-aware sequential network for capturing her time-sensitive demands. Besides retrieving relevant information from her personal history, we also propose to search and retrieve similar user's records as an additional reference. All these sequential records are further fused to make the final recommendation. Beyond this framework, we also develop a novel label trick that uses the previous labels (i.e., user's feedbacks) as the input to better capture the user's browsing pattern. We conduct extensive experiments on three real-world commercial datasets on click-through-rate prediction tasks against state-of-the-art methods. Experimental results demonstrate the superiority and efficiency of our proposed framework and techniques. Furthermore, results of online experiments on a daily item recommendation platform of Company X show that STARec gains average performance improvement of around 6% and 1.5% in its two main item recommendation scenarios on CTR metric respectively.

preprint2022arXiv

Model-based Multi-agent Policy Optimization with Adaptive Opponent-wise Rollouts

This paper investigates the model-based methods in multi-agent reinforcement learning (MARL). We specify the dynamics sample complexity and the opponent sample complexity in MARL, and conduct a theoretic analysis of return discrepancy upper bound. To reduce the upper bound with the intention of low sample complexity during the whole learning process, we propose a novel decentralized model-based MARL method, named Adaptive Opponent-wise Rollout Policy Optimization (AORPO). In AORPO, each agent builds its multi-agent environment model, consisting of a dynamics model and multiple opponent models, and trains its policy with the adaptive opponent-wise rollout. We further prove the theoretic convergence of AORPO under reasonable assumptions. Empirical experiments on competitive and cooperative tasks demonstrate that AORPO can achieve improved sample efficiency with comparable asymptotic performance over the compared MARL methods.

preprint2022arXiv

Multi-Level Interaction Reranking with User Behavior History

As the final stage of the multi-stage recommender system (MRS), reranking directly affects users' experience and satisfaction, thus playing a critical role in MRS. Despite the improvement achieved in the existing work, three issues are yet to be solved. First, users' historical behaviors contain rich preference information, such as users' long and short-term interests, but are not fully exploited in reranking. Previous work typically treats items in history equally important, neglecting the dynamic interaction between the history and candidate items. Second, existing reranking models focus on learning interactions at the item level while ignoring the fine-grained feature-level interactions. Lastly, estimating the reranking score on the ordered initial list before reranking may lead to the early scoring problem, thereby yielding suboptimal reranking performance. To address the above issues, we propose a framework named Multi-level Interaction Reranking (MIR). MIR combines low-level cross-item interaction and high-level set-to-list interaction, where we view the candidate items to be reranked as a set and the users' behavior history in chronological order as a list. We design a novel SLAttention structure for modeling the set-to-list interactions with personalized long-short term interests. Moreover, feature-level interactions are incorporated to capture the fine-grained influence among items. We design MIR in such a way that any permutation of the input items would not change the output ranking, and we theoretically prove it. Extensive experiments on three public and proprietary datasets show that MIR significantly outperforms the state-of-the-art models using various ranking and utility metrics.

preprint2022arXiv

Multi-Scale User Behavior Network for Entire Space Multi-Task Learning

Modelling the user's multiple behaviors is an essential part of modern e-commerce, whose widely adopted application is to jointly optimize click-through rate (CTR) and conversion rate (CVR) predictions. Most of existing methods overlook the effect of two key characteristics of the user's behaviors: for each item list, (i) contextual dependence refers to that the user's behaviors on any item are not purely determinated by the item itself but also are influenced by the user's previous behaviors (e.g., clicks, purchases) on other items in the same sequence; (ii) multiple time scales means that users are likely to click frequently but purchase periodically. To this end, we develop a new multi-scale user behavior network named Hierarchical rEcurrent Ranking On the Entire Space (HEROES) which incorporates the contextual information to estimate the user multiple behaviors in a multi-scale fashion. Concretely, we introduce a hierarchical framework, where the lower layer models the user's engagement behaviors while the upper layer estimates the user's satisfaction behaviors. The proposed architecture can automatically learn a suitable time scale for each layer to capture the dynamic user's behavioral patterns. Besides the architecture, we also introduce the Hawkes process to form a novel recurrent unit which can not only encode the items' features in the context but also formulate the excitation or discouragement from the user's previous behaviors. We further show that HEROES can be extended to build unbiased ranking systems through combinations with the survival analysis technique. Extensive experiments over three large-scale industrial datasets demonstrate the superiority of our model compared with the state-of-the-art methods.

preprint2022arXiv

Multi-View Graph Representation for Programming Language Processing: An Investigation into Algorithm Detection

Program representation, which aims at converting program source code into vectors with automatically extracted features, is a fundamental problem in programming language processing (PLP). Recent work tries to represent programs with neural networks based on source code structures. However, such methods often focus on the syntax and consider only one single perspective of programs, limiting the representation power of models. This paper proposes a multi-view graph (MVG) program representation method. MVG pays more attention to code semantics and simultaneously includes both data flow and control flow as multiple views. These views are then combined and processed by a graph neural network (GNN) to obtain a comprehensive program representation that covers various aspects. We thoroughly evaluate our proposed MVG approach in the context of algorithm detection, an important and challenging subfield of PLP. Specifically, we use a public dataset POJ-104 and also construct a new challenging dataset ALG-109 to test our method. In experiments, MVG outperforms previous methods significantly, demonstrating our model's strong capability of representing source code.

preprint2022arXiv

Neural Re-ranking in Multi-stage Recommender Systems: A Review

As the final stage of the multi-stage recommender system (MRS), re-ranking directly affects user experience and satisfaction by rearranging the input ranking lists, and thereby plays a critical role in MRS. With the advances in deep learning, neural re-ranking has become a trending topic and been widely applied in industrial applications. This review aims at integrating re-ranking algorithms into a broader picture, and paving ways for more comprehensive solutions for future research. For this purpose, we first present a taxonomy of current methods on neural re-ranking. Then we give a description of these methods along with the historic development according to their objectives. The network structure, personalization, and complexity are also discussed and compared. Next, we provide benchmarks of the major neural re-ranking models and quantitatively analyze their re-ranking performance. Finally, the review concludes with a discussion on future prospects of this field. A list of papers discussed in this review, the benchmark datasets, our re-ranking library LibRerank, and detailed parameter settings are publicly available at https://github.com/LibRerank-Community/LibRerank.

preprint2022arXiv

Offline Pre-trained Multi-Agent Decision Transformer: One Big Sequence Model Tackles All SMAC Tasks

Offline reinforcement learning leverages previously-collected offline datasets to learn optimal policies with no necessity to access the real environment. Such a paradigm is also desirable for multi-agent reinforcement learning (MARL) tasks, given the increased interactions among agents and with the enviroment. Yet, in MARL, the paradigm of offline pre-training with online fine-tuning has not been studied, nor datasets or benchmarks for offline MARL research are available. In this paper, we facilitate the research by providing large-scale datasets, and use them to examine the usage of the Decision Transformer in the context of MARL. We investigate the generalisation of MARL offline pre-training in the following three aspects: 1) between single agents and multiple agents, 2) from offline pretraining to the online fine-tuning, and 3) to that of multiple downstream tasks with few-shot and zero-shot capabilities. We start by introducing the first offline MARL dataset with diverse quality levels based on the StarCraftII environment, and then propose the novel architecture of multi-agent decision transformer (MADT) for effective offline learning. MADT leverages transformer's modelling ability of sequence modelling and integrates it seamlessly with both offline and online MARL tasks. A crucial benefit of MADT is that it learns generalisable policies that can transfer between different types of agents under different task scenarios. On StarCraft II offline dataset, MADT outperforms the state-of-the-art offline RL baselines. When applied to online tasks, the pre-trained MADT significantly improves sample efficiency, and enjoys strong performance both few-short and zero-shot cases. To our best knowledge, this is the first work that studies and demonstrates the effectiveness of offline pre-trained models in terms of sample efficiency and generalisability enhancements in MARL.

preprint2022arXiv

On Effective Scheduling of Model-based Reinforcement Learning

Model-based reinforcement learning has attracted wide attention due to its superior sample efficiency. Despite its impressive success so far, it is still unclear how to appropriately schedule the important hyperparameters to achieve adequate performance, such as the real data ratio for policy optimization in Dyna-style model-based algorithms. In this paper, we first theoretically analyze the role of real data in policy training, which suggests that gradually increasing the ratio of real data yields better performance. Inspired by the analysis, we propose a framework named AutoMBPO to automatically schedule the real data ratio as well as other hyperparameters in training model-based policy optimization (MBPO) algorithm, a representative running case of model-based methods. On several continuous control tasks, the MBPO instance trained with hyperparameters scheduled by AutoMBPO can significantly surpass the original one, and the real data ratio schedule found by AutoMBPO shows consistency with our theoretical analysis.

preprint2022arXiv

Spatio-Temporal Graph Few-Shot Learning with Cross-City Knowledge Transfer

Spatio-temporal graph learning is a key method for urban computing tasks, such as traffic flow, taxi demand and air quality forecasting. Due to the high cost of data collection, some developing cities have few available data, which makes it infeasible to train a well-performed model. To address this challenge, cross-city knowledge transfer has shown its promise, where the model learned from data-sufficient cities is leveraged to benefit the learning process of data-scarce cities. However, the spatio-temporal graphs among different cities show irregular structures and varied features, which limits the feasibility of existing Few-Shot Learning (\emph{FSL}) methods. Therefore, we propose a model-agnostic few-shot learning framework for spatio-temporal graph called ST-GFSL. Specifically, to enhance feature extraction by transfering cross-city knowledge, ST-GFSL proposes to generate non-shared parameters based on node-level meta knowledge. The nodes in target city transfer the knowledge via parameter matching, retrieving from similar spatio-temporal characteristics. Furthermore, we propose to reconstruct the graph structure during meta-learning. The graph reconstruction loss is defined to guide structure-aware learning, avoiding structure deviation among different datasets. We conduct comprehensive experiments on four traffic speed prediction benchmarks and the results demonstrate the effectiveness of ST-GFSL compared with state-of-the-art methods.

preprint2022arXiv

Towards Applicable Reinforcement Learning: Improving the Generalization and Sample Efficiency with Policy Ensemble

It is challenging for reinforcement learning (RL) algorithms to succeed in real-world applications like financial trading and logistic system due to the noisy observation and environment shifting between training and evaluation. Thus, it requires both high sample efficiency and generalization for resolving real-world tasks. However, directly applying typical RL algorithms can lead to poor performance in such scenarios. Considering the great performance of ensemble methods on both accuracy and generalization in supervised learning (SL), we design a robust and applicable method named Ensemble Proximal Policy Optimization (EPPO), which learns ensemble policies in an end-to-end manner. Notably, EPPO combines each policy and the policy ensemble organically and optimizes both simultaneously. In addition, EPPO adopts a diversity enhancement regularization over the policy space which helps to generalize to unseen states and promotes exploration. We theoretically prove EPPO increases exploration efficacy, and through comprehensive experimental evaluations on various tasks, we demonstrate that EPPO achieves higher efficiency and is robust for real-world applications compared with vanilla policy optimization algorithms and other ensemble methods. Code and supplemental materials are available at https://seqml.github.io/eppo.

preprint2022arXiv

Towards Collaborative Question Answering: A Preliminary Study

Knowledge and expertise in the real-world can be disjointedly owned. To solve a complex question, collaboration among experts is often called for. In this paper, we propose CollabQA, a novel QA task in which several expert agents coordinated by a moderator work together to answer questions that cannot be answered with any single agent alone. We make a synthetic dataset of a large knowledge graph that can be distributed to experts. We define the process to form a complex question from ground truth reasoning path, neural network agent models that can learn to solve the task, and evaluation metrics to check the performance. We show that the problem can be challenging without introducing prior of the collaboration structure, unless experts are perfect and uniform. Based on this experience, we elaborate extensions needed to approach collaboration tasks in real-world settings.

preprint2022arXiv

Towards Making the Most of BERT in Neural Machine Translation

GPT-2 and BERT demonstrate the effectiveness of using pre-trained language models (LMs) on various natural language processing tasks. However, LM fine-tuning often suffers from catastrophic forgetting when applied to resource-rich tasks. In this work, we introduce a concerted training framework (CTNMT) that is the key to integrate the pre-trained LMs to neural machine translation (NMT). Our proposed CTNMT consists of three techniques: a) asymptotic distillation to ensure that the NMT model can retain the previous pre-trained knowledge; b) a dynamic switching gate to avoid catastrophic forgetting of pre-trained knowledge; and c) a strategy to adjust the learning paces according to a scheduled policy. Our experiments in machine translation show CTNMT gains of up to 3 BLEU score on the WMT14 English-German language pair which even surpasses the previous state-of-the-art pre-training aided NMT by 1.4 BLEU score. While for the large WMT14 English-French task with 40 millions of sentence-pairs, our base model still significantly improves upon the state-of-the-art Transformer big model by more than 1 BLEU score. The code and model can be downloaded from https://github.com/bytedance/neurst/ tree/master/examples/ctnmt.

preprint2022arXiv

Towards Return Parity in Markov Decision Processes

Algorithmic decisions made by machine learning models in high-stakes domains may have lasting impacts over time. However, naive applications of standard fairness criterion in static settings over temporal domains may lead to delayed and adverse effects. To understand the dynamics of performance disparity, we study a fairness problem in Markov decision processes (MDPs). Specifically, we propose return parity, a fairness notion that requires MDPs from different demographic groups that share the same state and action spaces to achieve approximately the same expected time-discounted rewards. We first provide a decomposition theorem for return disparity, which decomposes the return disparity of any two MDPs sharing the same state and action spaces into the distance between group-wise reward functions, the discrepancy of group policies, and the discrepancy between state visitation distributions induced by the group policies. Motivated by our decomposition theorem, we propose algorithms to mitigate return disparity via learning a shared group policy with state visitation distributional alignment using integral probability metrics. We conduct experiments to corroborate our results, showing that the proposed algorithm can successfully close the disparity gap while maintaining the performance of policies on two real-world recommender system benchmark datasets.

preprint2022arXiv

Who to Watch Next: Two-side Interactive Networks for Live Broadcast Recommendation

With the prevalence of live broadcast business nowadays, a new type of recommendation service, called live broadcast recommendation, is widely used in many mobile e-commerce Apps. Different from classical item recommendation, live broadcast recommendation is to automatically recommend user anchors instead of items considering the interactions among triple-objects (i.e., users, anchors, items) rather than binary interactions between users and items. Existing methods based on binary objects, ranging from early matrix factorization to recently emerged deep learning, obtain objects' embeddings by mapping from pre-existing features. Directly applying these techniques would lead to limited performance, as they are failing to encode collaborative signals among triple-objects. In this paper, we propose a novel TWo-side Interactive NetworkS (TWINS) for live broadcast recommendation. In order to fully use both static and dynamic information on user and anchor sides, we combine a product-based neural network with a recurrent neural network to learn the embedding of each object. In addition, instead of directly measuring the similarity, TWINS effectively injects the collaborative effects into the embedding process in an explicit manner by modeling interactive patterns between the user's browsing history and the anchor's broadcast history in both item and anchor aspects. Furthermore, we design a novel co-retrieval technique to select key items among massive historic records efficiently. Offline experiments on real large-scale data show the superior performance of the proposed TWINS, compared to representative methods; and further results of online experiments on Diantao App show that TWINS gains average performance improvement of around 8% on ACTR metric, 3% on UCTR metric, 3.5% on UCVR metric.

preprint2021arXiv

Fork or Fail: Cycle-Consistent Training with Many-to-One Mappings

Cycle-consistent training is widely used for jointly learning a forward and inverse mapping between two domains of interest without the cumbersome requirement of collecting matched pairs within each domain. In this regard, the implicit assumption is that there exists (at least approximately) a ground-truth bijection such that a given input from either domain can be accurately reconstructed from successive application of the respective mappings. But in many applications no such bijection can be expected to exist and large reconstruction errors can compromise the success of cycle-consistent training. As one important instance of this limitation, we consider practically-relevant situations where there exists a many-to-one or surjective mapping between domains. To address this regime, we develop a conditional variational autoencoder (CVAE) approach that can be viewed as converting surjective mappings to implicit bijections whereby reconstruction errors in both directions can be minimized, and as a natural byproduct, realistic output diversity can be obtained in the one-to-many direction. As theoretical motivation, we analyze a simplified scenario whereby minima of the proposed CVAE-based energy function align with the recovery of ground-truth surjective mappings. On the empirical side, we consider a synthetic image dataset with known ground-truth, as well as a real-world application involving natural language generation from knowledge graphs and vice versa, a prototypical surjective case. For the latter, our CVAE pipeline can capture such many-to-one mappings during cycle training while promoting textural diversity for graph-to-text tasks. Our code is available at github.com/QipengGuo/CycleGT *A condensed version of this paper has been accepted to AISTATS 2021. This version contains additional content and updates.

preprint2021arXiv

NeoRL: A Near Real-World Benchmark for Offline Reinforcement Learning

Offline reinforcement learning (RL) aims at learning a good policy from a batch of collected data, without extra interactions with the environment during training. However, current offline RL benchmarks commonly have a large reality gap, because they involve large datasets collected by highly exploratory policies, and the trained policy is directly evaluated in the environment. In real-world situations, running a highly exploratory policy is prohibited to ensure system safety, the data is commonly very limited, and a trained policy should be well validated before deployment. In this paper, we present a near real-world offline RL benchmark, named NeoRL, which contains datasets from various domains with controlled sizes, and extra test datasets for policy validation. We evaluate existing offline RL algorithms on NeoRL and argue that the performance of a policy should also be compared with the deterministic version of the behavior policy, instead of the dataset reward. The empirical results demonstrate that the tested offline RL algorithms become less competitive to the deterministic policy on many datasets, and the offline policy evaluation hardly helps. The NeoRL suit can be found at http://polixir.ai/research/neorl. We hope this work will shed some light on future research and draw more attention when deploying RL in real-world systems.

preprint2021arXiv

Towards Generalized Implementation of Wasserstein Distance in GANs

Wasserstein GANs (WGANs), built upon the Kantorovich-Rubinstein (KR) duality of Wasserstein distance, is one of the most theoretically sound GAN models. However, in practice it does not always outperform other variants of GANs. This is mostly due to the imperfect implementation of the Lipschitz condition required by the KR duality. Extensive work has been done in the community with different implementations of the Lipschitz constraint, which, however, is still hard to satisfy the restriction perfectly in practice. In this paper, we argue that the strong Lipschitz constraint might be unnecessary for optimization. Instead, we take a step back and try to relax the Lipschitz constraint. Theoretically, we first demonstrate a more general dual form of the Wasserstein distance called the Sobolev duality, which relaxes the Lipschitz constraint but still maintains the favorable gradient property of the Wasserstein distance. Moreover, we show that the KR duality is actually a special case of the Sobolev duality. Based on the relaxed duality, we further propose a generalized WGAN training scheme named Sobolev Wasserstein GAN (SWGAN), and empirically demonstrate the improvement of SWGAN over existing methods with extensive experiments.

preprint2020arXiv

A Deep Recurrent Survival Model for Unbiased Ranking

Position bias is a critical problem in information retrieval when dealing with implicit yet biased user feedback data. Unbiased ranking methods typically rely on causality models and debias the user feedback through inverse propensity weighting. While practical, these methods still suffer from two major problems. First, when inferring a user click, the impact of the contextual information, such as documents that have been examined, is often ignored. Second, only the position bias is considered but other issues resulted from user browsing behaviors are overlooked. In this paper, we propose an end-to-end Deep Recurrent Survival Ranking (DRSR), a unified framework to jointly model user's various behaviors, to (i) consider the rich contextual information in the ranking list; and (ii) address the hidden issues underlying user behaviors, i.e., to mine observe pattern in queries without any click (non-click queries), and to model tracking logs which cannot truly reflect the user browsing intents (untrusted observation). Specifically, we adopt a recurrent neural network to model the contextual information and estimates the conditional likelihood of user feedback at each position. We then incorporate survival analysis techniques with the probability chain rule to mathematically recover the unbiased joint probability of one user's various behaviors. DRSR can be easily incorporated with both point-wise and pair-wise learning objectives. The extensive experiments over two large-scale industrial datasets demonstrate the significant performance gains of our model comparing with the state-of-the-arts.

preprint2020arXiv

An Efficient Neighborhood-based Interaction Model for Recommendation on Heterogeneous Graph

There is an influx of heterogeneous information network (HIN) based recommender systems in recent years since HIN is capable of characterizing complex graphs and contains rich semantics. Although the existing approaches have achieved performance improvement, while practical, they still face the following problems. On one hand, most existing HIN-based methods rely on explicit path reachability to leverage path-based semantic relatedness between users and items, e.g., metapath-based similarities. These methods are hard to use and integrate since path connections are sparse or noisy, and are often of different lengths. On the other hand, other graph-based methods aim to learn effective heterogeneous network representations by compressing node together with its neighborhood information into single embedding before prediction. This weakly coupled manner in modeling overlooks the rich interactions among nodes, which introduces an early summarization issue. In this paper, we propose an end-to-end Neighborhood-based Interaction Model for Recommendation (NIRec) to address the above problems. Specifically, we first analyze the significance of learning interactions in HINs and then propose a novel formulation to capture the interactive patterns between each pair of nodes through their metapath-guided neighborhoods. Then, to explore complex interactions between metapaths and deal with the learning complexity on large-scale networks, we formulate interaction in a convolutional way and learn efficiently with fast Fourier transform. The extensive experiments on four different types of heterogeneous graphs demonstrate the performance gains of NIRec comparing with state-of-the-arts. To the best of our knowledge, this is the first work providing an efficient neighborhood-based interaction model in the HIN-based recommendations.

preprint2020arXiv

An End-to-End Neighborhood-based Interaction Model for Knowledge-enhanced Recommendation

This paper studies graph-based recommendation, where an interaction graph is constructed from historical records and is lever-aged to alleviate data sparsity and cold start problems. We reveal an early summarization problem in existing graph-based models, and propose Neighborhood Interaction (NI) model to capture each neighbor pair (between user-side and item-side) distinctively. NI model is more expressive and can capture more complicated structural patterns behind user-item interactions. To further enrich node connectivity and utilize high-order structural information, we incorporate extra knowledge graphs (KGs) and adopt graph neural networks (GNNs) in NI, called Knowledge-enhanced NeighborhoodInteraction (KNI). Compared with the state-of-the-art recommendation methods,e.g., feature-based, meta path-based, and KG-based models, our KNI achieves superior performance in click-through rate prediction (1.1%-8.4% absolute AUC improvements) and out-performs by a wide margin in top-N recommendation on 4 real-world datasets.

preprint2020arXiv

Author Name Disambiguation on Heterogeneous Information Network with Adversarial Representation Learning

Author name ambiguity causes inadequacy and inconvenience in academic information retrieval, which raises the necessity of author name disambiguation (AND). Existing AND methods can be divided into two categories: the models focusing on content information to distinguish whether two papers are written by the same author, the models focusing on relation information to represent information as edges on the network and to quantify the similarity among papers. However, the former requires adequate labeled samples and informative negative samples, and are also ineffective in measuring the high-order connections among papers, while the latter needs complicated feature engineering or supervision to construct the network. We propose a novel generative adversarial framework to grow the two categories of models together: (i) the discriminative module distinguishes whether two papers are from the same author, and (ii) the generative module selects possibly homogeneous papers directly from the heterogeneous information network, which eliminates the complicated feature engineering. In such a way, the discriminative module guides the generative module to select homogeneous papers, and the generative module generates high-quality negative samples to train the discriminative module to make it aware of high-order connections among papers. Furthermore, a self-training strategy for the discriminative module and a random walk based generating algorithm are designed to make the training stable and efficient. Extensive experiments on two real-world AND benchmarks demonstrate that our model provides significant performance improvement over the state-of-the-art methods.

preprint2020arXiv

AutoFIS: Automatic Feature Interaction Selection in Factorization Models for Click-Through Rate Prediction

Learning feature interactions is crucial for click-through rate (CTR) prediction in recommender systems. In most existing deep learning models, feature interactions are either manually designed or simply enumerated. However, enumerating all feature interactions brings large memory and computation cost. Even worse, useless interactions may introduce noise and complicate the training process. In this work, we propose a two-stage algorithm called Automatic Feature Interaction Selection (AutoFIS). AutoFIS can automatically identify important feature interactions for factorization models with computational cost just equivalent to training the target model to convergence. In the \emph{search stage}, instead of searching over a discrete set of candidate feature interactions, we relax the choices to be continuous by introducing the architecture parameters. By implementing a regularized optimizer over the architecture parameters, the model can automatically identify and remove the redundant feature interactions during the training process of the model. In the \emph{re-train stage}, we keep the architecture parameters serving as an attention unit to further boost the performance. Offline experiments on three large-scale datasets (two public benchmarks, one private) demonstrate that AutoFIS can significantly improve various FM based models. AutoFIS has been deployed onto the training platform of Huawei App Store recommendation service, where a 10-day online A/B test demonstrated that AutoFIS improved the DeepFM model by 20.3\% and 20.1\% in terms of CTR and CVR respectively.

preprint2020arXiv

Bi-level Actor-Critic for Multi-agent Coordination

Coordination is one of the essential problems in multi-agent systems. Typically multi-agent reinforcement learning (MARL) methods treat agents equally and the goal is to solve the Markov game to an arbitrary Nash equilibrium (NE) when multiple equilibra exist, thus lacking a solution for NE selection. In this paper, we treat agents \emph{unequally} and consider Stackelberg equilibrium as a potentially better convergence point than Nash equilibrium in terms of Pareto superiority, especially in cooperative environments. Under Markov games, we formally define the bi-level reinforcement learning problem in finding Stackelberg equilibrium. We propose a novel bi-level actor-critic learning method that allows agents to have different knowledge base (thus intelligent), while their actions still can be executed simultaneously and distributedly. The convergence proof is given, while the resulting learning algorithm is tested against the state of the arts. We found that the proposed bi-level actor-critic algorithm successfully converged to the Stackelberg equilibria in matrix games and find an asymmetric solution in a highway merge environment.

preprint2020arXiv

GeneraLight: Improving Environment Generalization of Traffic Signal Control via Meta Reinforcement Learning

The heavy traffic congestion problem has always been a concern for modern cities. To alleviate traffic congestion, researchers use reinforcement learning (RL) to develop better traffic signal control (TSC) algorithms in recent years. However, most RL models are trained and tested in the same traffic flow environment, which results in a serious overfitting problem. Since the traffic flow environment in the real world keeps varying, these models can hardly be applied due to the lack of generalization ability. Besides, the limited number of accessible traffic flow data brings extra difficulty in testing the generalization ability of the models. In this paper, we design a novel traffic flow generator based on Wasserstein generative adversarial network to generate sufficient diverse and quality traffic flows and use them to build proper training and testing environments. Then we propose a meta-RL TSC framework GeneraLight to improve the generalization ability of TSC models. GeneraLight boosts the generalization performance by combining the idea of flow clustering and model-agnostic meta-learning. We conduct extensive experiments on multiple real-world datasets to show the superior performance of GeneraLight on generalizing to different traffic flows.

preprint2020arXiv

GIKT: A Graph-based Interaction Model for Knowledge Tracing

With the rapid development in online education, knowledge tracing (KT) has become a fundamental problem which traces students' knowledge status and predicts their performance on new questions. Questions are often numerous in online education systems, and are always associated with much fewer skills. However, the previous literature fails to involve question information together with high-order question-skill correlations, which is mostly limited by data sparsity and multi-skill problems. From the model perspective, previous models can hardly capture the long-term dependency of student exercise history, and cannot model the interactions between student-questions, and student-skills in a consistent way. In this paper, we propose a Graph-based Interaction model for Knowledge Tracing (GIKT) to tackle the above probems. More specifically, GIKT utilizes graph convolutional network (GCN) to substantially incorporate question-skill correlations via embedding propagation. Besides, considering that relevant questions are usually scattered throughout the exercise history, and that question and skill are just different instantiations of knowledge, GIKT generalizes the degree of students' master of the question to the interactions between the student's current state, the student's history related exercises, the target question, and related skills. Experiments on three datasets demonstrate that GIKT achieves the new state-of-the-art performance, with at least 1% absolute AUC improvement.

preprint2020arXiv

GraphAF: a Flow-based Autoregressive Model for Molecular Graph Generation

Molecular graph generation is a fundamental problem for drug discovery and has been attracting growing attention. The problem is challenging since it requires not only generating chemically valid molecular structures but also optimizing their chemical properties in the meantime. Inspired by the recent progress in deep generative models, in this paper we propose a flow-based autoregressive model for graph generation called GraphAF. GraphAF combines the advantages of both autoregressive and flow-based approaches and enjoys: (1) high model flexibility for data density estimation; (2) efficient parallel computation for training; (3) an iterative sampling process, which allows leveraging chemical domain knowledge for valency checking. Experimental results show that GraphAF is able to generate 68% chemically valid molecules even without chemical knowledge rules and 100% valid molecules with chemical rules. The training process of GraphAF is two times faster than the existing state-of-the-art approach GCPN. After fine-tuning the model for goal-directed property optimization with reinforcement learning, GraphAF achieves state-of-the-art performance on both chemical property optimization and constrained property optimization.

preprint2020arXiv

Interactive Recommender System via Knowledge Graph-enhanced Reinforcement Learning

Interactive recommender system (IRS) has drawn huge attention because of its flexible recommendation strategy and the consideration of optimal long-term user experiences. To deal with the dynamic user preference and optimize accumulative utilities, researchers have introduced reinforcement learning (RL) into IRS. However, RL methods share a common issue of sample efficiency, i.e., huge amount of interaction data is required to train an effective recommendation policy, which is caused by the sparse user responses and the large action space consisting of a large number of candidate items. Moreover, it is infeasible to collect much data with explorative policies in online environments, which will probably harm user experience. In this work, we investigate the potential of leveraging knowledge graph (KG) in dealing with these issues of RL methods for IRS, which provides rich side information for recommendation decision making. Instead of learning RL policies from scratch, we make use of the prior knowledge of the item correlation learned from KG to (i) guide the candidate selection for better candidate item retrieval, (ii) enrich the representation of items and user states, and (iii) propagate user preferences among the correlated items over KG to deal with the sparsity of user feedback. Comprehensive experiments have been conducted on two real-world datasets, which demonstrate the superiority of our approach with significant improvements against state-of-the-arts.

preprint2020arXiv

Large-Scale Optimal Transport via Adversarial Training with Cycle-Consistency

Recent advances in large-scale optimal transport have greatly extended its application scenarios in machine learning. However, existing methods either not explicitly learn the transport map or do not support general cost function. In this paper, we propose an end-to-end approach for large-scale optimal transport, which directly solves the transport map and is compatible with general cost function. It models the transport map via stochastic neural networks and enforces the constraint on the marginal distributions via adversarial training. The proposed framework can be further extended towards learning Monge map or optimal bijection via adopting cycle-consistency constraint(s). We verify the effectiveness of the proposed method and demonstrate its superior performance against existing methods with large-scale real-world applications, including domain adaptation, image-to-image translation, and color transfer.

preprint2020arXiv

Learning to Infer User Hidden States for Online Sequential Advertising

To drive purchase in online advertising, it is of the advertiser's great interest to optimize the sequential advertising strategy whose performance and interpretability are both important. The lack of interpretability in existing deep reinforcement learning methods makes it not easy to understand, diagnose and further optimize the strategy. In this paper, we propose our Deep Intents Sequential Advertising (DISA) method to address these issues. The key part of interpretability is to understand a consumer's purchase intent which is, however, unobservable (called hidden states). In this paper, we model this intention as a latent variable and formulate the problem as a Partially Observable Markov Decision Process (POMDP) where the underlying intents are inferred based on the observable behaviors. Large-scale industrial offline and online experiments demonstrate our method's superior performance over several baselines. The inferred hidden states are analyzed, and the results prove the rationality of our inference.

preprint2020arXiv

Multi-Agent Determinantal Q-Learning

Centralized training with decentralized execution has become an important paradigm in multi-agent learning. Though practical, current methods rely on restrictive assumptions to decompose the centralized value function across agents for execution. In this paper, we eliminate this restriction by proposing multi-agent determinantal Q-learning. Our method is established on Q-DPP, an extension of determinantal point process (DPP) with partition-matroid constraint to multi-agent setting. Q-DPP promotes agents to acquire diverse behavioral models; this allows a natural factorization of the joint Q-functions with no need for \emph{a priori} structural constraints on the value function or special network architectures. We demonstrate that Q-DPP generalizes major solutions including VDN, QMIX, and QTRAN on decentralizable cooperative tasks. To efficiently draw samples from Q-DPP, we adopt an existing sample-by-projection sampler with theoretical approximation guarantee. The sampler also benefits exploration by coordinating agents to cover orthogonal directions in the state space during multi-agent training. We evaluate our algorithm on various cooperative benchmarks; its effectiveness has been demonstrated when compared with the state-of-the-art.

preprint2020arXiv

Multi-Agent Interactions Modeling with Correlated Policies

In multi-agent systems, complex interacting behaviors arise due to the high correlations among agents. However, previous work on modeling multi-agent interactions from demonstrations is primarily constrained by assuming the independence among policies and their reward structures. In this paper, we cast the multi-agent interactions modeling problem into a multi-agent imitation learning framework with explicit modeling of correlated policies by approximating opponents' policies, which can recover agents' policies that can regenerate similar interactions. Consequently, we develop a Decentralized Adversarial Imitation Learning algorithm with Correlated policies (CoDAIL), which allows for decentralized training and execution. Various experiments demonstrate that CoDAIL can better regenerate complex interactions close to the demonstrators and outperforms state-of-the-art multi-agent imitation learning methods. Our code is available at \url{https://github.com/apexrl/CoDAIL}.

preprint2020arXiv

User Behavior Retrieval for Click-Through Rate Prediction

Click-through rate (CTR) prediction plays a key role in modern online personalization services. In practice, it is necessary to capture user's drifting interests by modeling sequential user behaviors to build an accurate CTR prediction model. However, as the users accumulate more and more behavioral data on the platforms, it becomes non-trivial for the sequential models to make use of the whole behavior history of each user. First, directly feeding the long behavior sequence will make online inference time and system load infeasible. Second, there is much noise in such long histories to fail the sequential model learning. The current industrial solutions mainly truncate the sequences and just feed recent behaviors to the prediction model, which leads to a problem that sequential patterns such as periodicity or long-term dependency are not embedded in the recent several behaviors but in far back history. To tackle these issues, in this paper we consider it from the data perspective instead of just designing more sophisticated yet complicated models and propose User Behavior Retrieval for CTR prediction (UBR4CTR) framework. In UBR4CTR, the most relevant and appropriate user behaviors will be firstly retrieved from the entire user history sequence using a learnable search method. These retrieved behaviors are then fed into a deep model to make the final prediction instead of simply using the most recent ones. It is highly feasible to deploy UBR4CTR into industrial model pipeline with low cost. Experiments on three real-world large-scale datasets demonstrate the superiority and efficacy of our proposed framework and models.

Institution

Affiliation not imported yet

This author record came from a source that does not expose affiliation metadata. Once the author claims the profile or we enrich the record from another provider, this section will link to the concrete institution.

Topic footprint

Fields this researcher appears in

Source provenance

Where this author record came from

arxivconfidence 95%

external id: arxiv:2605.12058:author:10:weinan-zhang

Imported May 20, 2026Synced May 21, 2026

arxivconfidence 95%

external id: arxiv:2604.27472:author:10:weinan-zhang