BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Jing Ma

Jing Ma contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
Artificial IntelligenceComputation and LanguageMachine Learningmath.RTInformation RetrievalComputer Visioncond-mat.mes-hallcond-mat.mtrl-scimath.NTphysics.app-ph

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
19works
0followers
10topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

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

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

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.

Published work

19 published item(s)

preprint2026arXiv

MemRouter: Memory-as-Embedding Routing for Long-Term Conversational Agents

Long-term conversational agents must decide which turns to store in external memory, yet recent systems rely on autoregressive LLM generation at every turn to make that decision. We present MemRouter, a write-side memory router that decouples memory admission from the downstream answer backbone and replaces per-turn memory-management decoding with an embedding-based routing policy. MemRouter encodes each turn together with recent context, projects the resulting embeddings through a frozen LLM backbone, and predicts whether the turn should be stored using lightweight classification heads while training only 12M parameters. Under a controlled matched-harness comparison on LoCoMo, where the retrieval pipeline, answer prompts, and QA backbone (Qwen2.5-7B) are held identical, MemRouter outperforms an LLM-based memory manager on every question category (overall F1 52.0 vs 45.6, non-overlapping 95% CIs) while reducing memory-management p50 latency from 970ms to 58ms. Descriptive factorial averaging further shows that learned admission improves mean F1 by +10.3 over random storage, category-specific prompting adds +5.2 over a generic prompt, and retrieval contributes +0.7. These results suggest that write-side memory admission can be learned by a small supervised router, while answer generation remains a separate downstream component in long-horizon conversational QA.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Simply Stabilizing the Loop via Fully Looped Transformer

Scaling model performance typically requires increasing model size. Looped Transformer offers a compelling alternative by iteratively reusing the same Transformer blocks, trading additional computation for improved performance without increasing parameter count or context length. Because the number of loop iterations can be adjusted at inference, it also provides a natural mechanism for balancing performance and test-time compute. However, Looped Transformer still suffers from training instability when the number of loop iterations increases. Our analysis reveals that this instability stems from two sources: gradient oscillation and residual explosion. To address these two problems, we propose the Fully Looped Transformer, which introduces two parameter-free modifications: (1) Fully Looped Architecture, which distributes inter-loop signals across all layers to mitigate residual explosion; (2) Attention Injection, which reuses the existing attention block to suppress gradient oscillation. These modifications stabilize training dynamics, enabling the Fully Looped Transformer to be trained stably up to 12 loop iterations, whereas other baseline looped models collapse in this regime. In milder settings where Looped Transformer does not collapse, Fully Looped Transformer still improves average downstream-task performance by up to 13.2\%. Overall, our experiments demonstrate that Fully Looped Transformer improves training stability, enhances downstream performance, and provides preliminary adaptability under different test-time compute budgets by varying loop iterations at inference.

0 citations0 reviewsNo code linkedOpen access
preprint2023arXiv

Causal Inference in Recommender Systems: A Survey of Strategies for Bias Mitigation, Explanation, and Generalization

In the era of information overload, recommender systems (RSs) have become an indispensable part of online service platforms. Traditional RSs estimate user interests and predict their future behaviors by utilizing correlations in the observational historical activities, their profiles, and the content of interacted items. However, since the inherent causal reasons that lead to the observed users' behaviors are not considered, multiple types of biases could exist in the generated recommendations. In addition, the causal motives that drive user activities are usually entangled in these RSs, where the explainability and generalization abilities of recommendations cannot be guaranteed. To address these drawbacks, recent years have witnessed an upsurge of interest in enhancing traditional RSs with causal inference techniques. In this survey, we provide a systematic overview of causal RSs and help readers gain a comprehensive understanding of this promising area. We start with the basic concepts of traditional RSs and their limitations due to the lack of causal reasoning ability. We then discuss how different causal inference techniques can be introduced to address these challenges, with an emphasis on debiasing, explainability promotion, and generalization improvement. Furthermore, we thoroughly analyze various evaluation strategies for causal RSs, focusing especially on how to reliably estimate their performance with biased data if the causal effects of interests are unavailable. Finally, we provide insights into potential directions for future causal RS research.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

(Gorenstein) silting modules in recollements

In the paper, we focus on the silting properties and the combinatorial properties of silting and Gorenstein, which is called Gorenstein silting, where the main tools used are recollements of module categories and tensor products. For a ring A and its idempotent ideal J, we show that an A/J-module T is a silting A-module if and only if T is a silting A/J-module. For the finite dimensional k-algebras, with k a field, we show that the tensor products of silting modules are still silting. We also show that the (partial) Gorenstein silting properties can be glued by the recollements of module categories of Noetherian rings. As a consequence, we glue the Gorenstein silting modules of an upper triangular matrix Gorenstein ring by those of the involved rings.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

A Frustratingly Simple Approach for End-to-End Image Captioning

Image Captioning is a fundamental task to join vision and language, concerning about cross-modal understanding and text generation. Recent years witness the emerging attention on image captioning. Most of existing works follow a traditional two-stage training paradigm. Before training the captioning models, an extra object detector is utilized to recognize the objects in the image at first. However, they require sizeable datasets with fine-grained object annotation for training the object detector, which is a daunting task. In addition, the errors of the object detectors are easy to propagate to the following captioning models, degenerating models' performance. To alleviate such defects, we propose a frustratingly simple but highly effective end-to-end image captioning framework, Visual Conditioned GPT (VC-GPT), by connecting the pre-trained visual encoder (CLIP-ViT) and language decoder (GPT2). Different from the vanilla connection method that directly inserts the cross-attention modules into GPT2, we come up with a self-ensemble cross-modal fusion mechanism that comprehensively considers both the single- and cross-modal knowledge. As a result, we do not need extra object detectors for model training. Experimental results conducted on three popular image captioning benchmarks (MSCOCO, Flickr30k and NoCaps) demonstrate that our VC-GPT achieves either the best or the second-best performance across all evaluation metrics over extensive baseline systems.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

A Weakly Supervised Propagation Model for Rumor Verification and Stance Detection with Multiple Instance Learning

The diffusion of rumors on microblogs generally follows a propagation tree structure, that provides valuable clues on how an original message is transmitted and responded by users over time. Recent studies reveal that rumor detection and stance detection are two different but relevant tasks which can jointly enhance each other, e.g., rumors can be debunked by cross-checking the stances conveyed by their relevant microblog posts, and stances are also conditioned on the nature of the rumor. However, most stance detection methods require enormous post-level stance labels for training, which are labor-intensive given a large number of posts. Enlightened by Multiple Instance Learning (MIL) scheme, we first represent the diffusion of claims with bottom-up and top-down trees, then propose two tree-structured weakly supervised frameworks to jointly classify rumors and stances, where only the bag-level labels concerning claim's veracity are needed. Specifically, we convert the multi-class problem into a multiple MIL-based binary classification problem where each binary model focuses on differentiating a target stance or rumor type and other types. Finally, we propose a hierarchical attention mechanism to aggregate the binary predictions, including (1) a bottom-up or top-down tree attention layer to aggregate binary stances into binary veracity; and (2) a discriminative attention layer to aggregate the binary class into finer-grained classes. Extensive experiments conducted on three Twitter-based datasets demonstrate promising performance of our model on both claim-level rumor detection and post-level stance classification compared with state-of-the-art methods.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

DecBERT: Enhancing the Language Understanding of BERT with Causal Attention Masks

Since 2017, the Transformer-based models play critical roles in various downstream Natural Language Processing tasks. However, a common limitation of the attention mechanism utilized in Transformer Encoder is that it cannot automatically capture the information of word order, so explicit position embeddings are generally required to be fed into the target model. In contrast, Transformer Decoder with the causal attention masks is naturally sensitive to the word order. In this work, we focus on improving the position encoding ability of BERT with the causal attention masks. Furthermore, we propose a new pre-trained language model DecBERT and evaluate it on the GLUE benchmark. Experimental results show that (1) the causal attention mask is effective for BERT on the language understanding tasks; (2) our DecBERT model without position embeddings achieve comparable performance on the GLUE benchmark; and (3) our modification accelerates the pre-training process and DecBERT w/ PE achieves better overall performance than the baseline systems when pre-training with the same amount of computational resources.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Detect Rumors in Microblog Posts for Low-Resource Domains via Adversarial Contrastive Learning

Massive false rumors emerging along with breaking news or trending topics severely hinder the truth. Existing rumor detection approaches achieve promising performance on the yesterday's news, since there is enough corpus collected from the same domain for model training. However, they are poor at detecting rumors about unforeseen events especially those propagated in different languages due to the lack of training data and prior knowledge (i.e., low-resource regimes). In this paper, we propose an adversarial contrastive learning framework to detect rumors by adapting the features learned from well-resourced rumor data to that of the low-resourced. Our model explicitly overcomes the restriction of domain and/or language usage via language alignment and a novel supervised contrastive training paradigm. Moreover, we develop an adversarial augmentation mechanism to further enhance the robustness of low-resource rumor representation. Extensive experiments conducted on two low-resource datasets collected from real-world microblog platforms demonstrate that our framework achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Empowering Next POI Recommendation with Multi-Relational Modeling

With the wide adoption of mobile devices and web applications, location-based social networks (LBSNs) offer large-scale individual-level location-related activities and experiences. Next point-of-interest (POI) recommendation is one of the most important tasks in LBSNs, aiming to make personalized recommendations of next suitable locations to users by discovering preferences from users' historical activities. Noticeably, LBSNs have offered unparalleled access to abundant heterogeneous relational information about users and POIs (including user-user social relations, such as families or colleagues; and user-POI visiting relations). Such relational information holds great potential to facilitate the next POI recommendation. However, most existing methods either focus on merely the user-POI visits, or handle different relations based on over-simplified assumptions while neglecting relational heterogeneities. To fill these critical voids, we propose a novel framework, MEMO, which effectively utilizes the heterogeneous relations with a multi-network representation learning module, and explicitly incorporates the inter-temporal user-POI mutual influence with the coupled recurrent neural networks. Extensive experiments on real-world LBSN data validate the superiority of our framework over the state-of-the-art next POI recommendation methods.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

From Gorenstein derived equivalences to stable functors of Gorenstein projective modules

In the paper, we mainly connect the Gorenstein derived equivalence and stable functors of Gorenstein projective modules. Specially, we prove that a Gorenstein derived equivalence between CM-finite algebras A and B can induce a stable functor between the factor categories A-mod/A-Gproj and B-mod\B-Gproj. Furthermore, the above stable functor is an equivalence when A and B are Gorenstein.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Gorenstein silting modules and Gorenstein projective modules

(Partial) Gorenstein silting modules are introduced and investigated. It is shown that for finite dimensional algebras of finite CM-type, partial Gorenstein silting modules are in bijection with τ_G-rigid modules; Gorenstein silting modules are the module-theoretic counterpart of 2-term Gorenstein silting complexes; and the relation between 2-term Gorenstein silting complexes, t-structures and torsion pairs in module categories. Furthermore, the corresponding version of the classical Brenner-Butler theorem in this setting are characterised; and the upper bound of the global dimension of endomorphism algebras of 2-term Gorenstein silting complexes over an algebra A are also characterised by terms of the Gorenstein global dimension of A.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Learning Causal Effects on Hypergraphs

Hypergraphs provide an effective abstraction for modeling multi-way group interactions among nodes, where each hyperedge can connect any number of nodes. Different from most existing studies which leverage statistical dependencies, we study hypergraphs from the perspective of causality. Specifically, in this paper, we focus on the problem of individual treatment effect (ITE) estimation on hypergraphs, aiming to estimate how much an intervention (e.g., wearing face covering) would causally affect an outcome (e.g., COVID-19 infection) of each individual node. Existing works on ITE estimation either assume that the outcome on one individual should not be influenced by the treatment assignments on other individuals (i.e., no interference), or assume the interference only exists between pairs of connected individuals in an ordinary graph. We argue that these assumptions can be unrealistic on real-world hypergraphs, where higher-order interference can affect the ultimate ITE estimations due to the presence of group interactions. In this work, we investigate high-order interference modeling, and propose a new causality learning framework powered by hypergraph neural networks. Extensive experiments on real-world hypergraphs verify the superiority of our framework over existing baselines.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Learning Fair Node Representations with Graph Counterfactual Fairness

Fair machine learning aims to mitigate the biases of model predictions against certain subpopulations regarding sensitive attributes such as race and gender. Among the many existing fairness notions, counterfactual fairness measures the model fairness from a causal perspective by comparing the predictions of each individual from the original data and the counterfactuals. In counterfactuals, the sensitive attribute values of this individual had been modified. Recently, a few works extend counterfactual fairness to graph data, but most of them neglect the following facts that can lead to biases: 1) the sensitive attributes of each node's neighbors may causally affect the prediction w.r.t. this node; 2) the sensitive attributes may causally affect other features and the graph structure. To tackle these issues, in this paper, we propose a novel fairness notion - graph counterfactual fairness, which considers the biases led by the above facts. To learn node representations towards graph counterfactual fairness, we propose a novel framework based on counterfactual data augmentation. In this framework, we generate counterfactuals corresponding to perturbations on each node's and their neighbors' sensitive attributes. Then we enforce fairness by minimizing the discrepancy between the representations learned from the original graph and the counterfactuals for each node. Experiments on both synthetic and real-world graphs show that our framework outperforms the state-of-the-art baselines in graph counterfactual fairness, and also achieves comparable prediction performance.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Lifting of recollements and Gorenstein projective modules

In the paper, we investigate the lifting of recollements with respect to Gorenstein-projective modules. Specifically, a homological ring epimorphism can induce a lifting of the recollement of the stable category of finitely generated Gorenstein-projective modules; the recollement of the bounded Gorenstein derived categories of some upper triangular matrix algebras can be lifted to the homotopy category of Gorenstein-projective modules. As a byproduct, we give a sufficient and necessary condition on the upper triangular matrix algebra T_{n}(A) to be of finite CM-type for an algebra A of finite CM-type.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Recycling of Perovskite Substrate

The use of water-soluble sacrificial layer of Sr$_3$Al$_2$O$_6$ has tremendously boosted the research on freestanding functional oxide thin films, especially thanks to its ultimate capability to produce high-quality epitaxial perovskite thin films. However, the costly single-crystalline substrates, e.g. SrTiO$_3$, were generally discarded after obtaining the freestanding thin films. Here, we demonstrate that the SrTiO$_3$ substrates can be recycled to fabricate La$_{0.7}$Sr$_{0.3}$MnO$_3$ films with nearly identical structural and electrical properties. After attaining freestanding thin films, the residues on SrTiO$_3$ can be removed by 80 \degree C hot water soaking and rinsing treatments. Consequently, the surface of SrTiO$_3$ reverted to its original step-and-terrace structure.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Review-guided Helpful Answer Identification in E-commerce

Product-specific community question answering platforms can greatly help address the concerns of potential customers. However, the user-provided answers on such platforms often vary a lot in their qualities. Helpfulness votes from the community can indicate the overall quality of the answer, but they are often missing. Accurately predicting the helpfulness of an answer to a given question and thus identifying helpful answers is becoming a demanding need. Since the helpfulness of an answer depends on multiple perspectives instead of only topical relevance investigated in typical QA tasks, common answer selection algorithms are insufficient for tackling this task. In this paper, we propose the Review-guided Answer Helpfulness Prediction (RAHP) model that not only considers the interactions between QA pairs but also investigates the opinion coherence between the answer and crowds' opinions reflected in the reviews, which is another important factor to identify helpful answers. Moreover, we tackle the task of determining opinion coherence as a language inference problem and explore the utilization of pre-training strategy to transfer the textual inference knowledge obtained from a specifically designed trained network. Extensive experiments conducted on real-world data across seven product categories show that our proposed model achieves superior performance on the prediction task.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Spatio-Temporal Tensor Sketching via Adaptive Sampling

Mining massive spatio-temporal data can help a variety of real-world applications such as city capacity planning, event management, and social network analysis. The tensor representation can be used to capture the correlation between space and time and simultaneously exploit the latent structure of the spatial and temporal patterns in an unsupervised fashion. However, the increasing volume of spatio-temporal data has made it prohibitively expensive to store and analyze using tensor factorization. In this paper, we propose SkeTenSmooth, a novel tensor factorization framework that uses adaptive sampling to compress the tensor in a temporally streaming fashion and preserves the underlying global structure. SkeTenSmooth adaptively samples incoming tensor slices according to the detected data dynamics. Thus, the sketches are more representative and informative of the tensor dynamic patterns. In addition, we propose a robust tensor factorization method that can deal with the sketched tensor and recover the original patterns. Experiments on the New York City Yellow Taxi data show that SkeTenSmooth greatly reduces the memory cost and outperforms random sampling and fixed rate sampling method in terms of retaining the underlying patterns.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Gate-tunable flat bands in van der Waals patterned dielectric superlattices

Superlattice engineering provides the means to reshape the fabric felt by quasiparticles moving in a material. Here we argue that bandstructure engineering with superlattices can be pushed to the extreme limit by stacking gapped van der Waals (vdW) materials on patterned dielectric substrates. Specifically, we find that high quality vdW patterned dielectric superlattices (PDS) realize a series of robust flat bands that can be directly switched on and off by gate voltage in situ. In contrast to existing superlattice platforms, these flat bands are realized without the need for fine tuning. Instead, the bands become flat as the gate voltage increases in magnitude. The characteristics of PDS flatbands are highly tunable: the type of flatband (single non-degenerate or dirac-cone-like), localization length, and interaction energy are sensitive to the applied gate voltage. As a result, electron-electron interactions in the PDS flatbands can become stronger than both the bandwidth and disorder broadening, providing a setting for correlated behavior such as flatband ferromagnetism. We expect PDS flatbands can be experimentally realized in a range of readily available gapped vdW materials such as monolayer transition metal dichalcogenides, e.g. WSe2.

0 citations0 reviewsNo code linkedOpen access