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

Xiaodong Liu

Xiaodong Liu contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
Computation and LanguageMachine LearningArtificial Intelligenceastro-ph.EPmath.APComputer VisionComputational Engineering, Finance, and Sciencecond-mat.dis-nneess.IVeess.SPInformation Theorymath.ITmath.NANeural and Evolutionary ComputingNumerical Analysisphysics.comp-phphysics.flu-dynphysics.space-phSoftware Engineering

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
34works
0followers
19topics
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

34 published item(s)

preprint2026arXiv

Cubit: Token Mixer with Kernel Ridge Regression

Since its introduction in 2017, the Transformer has become one of the most widely adopted architectures in modern deep learning. Despite extensive efforts to improve positional encoding, attention mechanisms, and feed-forward networks, the core token-mixing mechanism in Transformers remains attention. In this work, we show that the attention module in Transformers can be interpreted as performing Nadaraya-Watson regression, where it computes similarities between tokens and aggregates the corresponding values accordingly. Motivated by this perspective, we propose Cubit, a potential next-generation architecture that leverages Kernel Ridge Regression (KRR), while the vanilla Transformer relies on Nadaraya-Watson regression. Specifically, Cubit modifies the classical attention computation by incorporating the closed-form solution of KRR, combining value aggregation through kernel similarities with normalization via the inverse of the kernel matrix. To improve the training stability, we further propose the Limited-Range Rescale (LRR), which rescales the value layer within a controlled range. We argue that Cubit, as a KRR-based architecture, provides a stronger mathematical foundation than the vanilla Transformer, whose attention mechanism corresponds to Nadaraya-Watson regression. We validate this claim through comprehensive experiments. The experimental results suggest that Cubit may exhibit stronger long-sequence modeling capability. In particular, its performance gain over the Transformer appears to increase as the training sequence length grows.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

General reasoning represents a long-standing and formidable challenge in artificial intelligence. Recent breakthroughs, exemplified by large language models (LLMs) and chain-of-thought prompting, have achieved considerable success on foundational reasoning tasks. However, this success is heavily contingent upon extensive human-annotated demonstrations, and models' capabilities are still insufficient for more complex problems. Here we show that the reasoning abilities of LLMs can be incentivized through pure reinforcement learning (RL), obviating the need for human-labeled reasoning trajectories. The proposed RL framework facilitates the emergent development of advanced reasoning patterns, such as self-reflection, verification, and dynamic strategy adaptation. Consequently, the trained model achieves superior performance on verifiable tasks such as mathematics, coding competitions, and STEM fields, surpassing its counterparts trained via conventional supervised learning on human demonstrations. Moreover, the emergent reasoning patterns exhibited by these large-scale models can be systematically harnessed to guide and enhance the reasoning capabilities of smaller models.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

EMP: Enhance Memory in Data Pruning

Recently, large language and vision models have shown strong performance, but due to high pre-training and fine-tuning costs, research has shifted towards faster training via dataset pruning. Previous methods used sample loss as an evaluation criterion, aiming to select the most "difficult" samples for training. However, when the pruning rate increases, the number of times each sample is trained becomes more evenly distributed, which causes many critical or general samples to not be effectively fitted. We refer to this as Low-Frequency Learning (LFL). In other words, LFL prevents the model from remembering most samples. In our work, we decompose the scoring function of LFL, provide a theoretical explanation for the inefficiency of LFL, and propose adding a memory term to the scoring function to enhance the model's memory capability, along with an approximation of this memory term. Similarly, we explore memory in Self-Supervised Learning (SSL), marking the first discussion on SSL memory. Using contrastive learning, we derive the memory term both theoretically and experimentally. Finally, we propose Enhance Memory Pruning (EMP), which addresses the issue of insufficient memory under high pruning rates by enhancing the model's memory of data, thereby improving its performance. We evaluated the performance of EMP in tasks such as image classification, natural language understanding, and model pre-training. The results show that EMP can improve model performance under extreme pruning rates. For example, in the CIFAR100-ResNet50 pre-training task, with 70\% pruning, EMP outperforms current methods by 2.2\%.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

MultiBreak: A Scalable and Diverse Multi-turn Jailbreak Benchmark for Evaluating LLM Safety

We present MultiBreak, a scalable and diverse multi-turn jailbreak benchmark to evaluate large language model (LLM) safety. Multi-turn jailbreaks mimic natural conversational settings, making them easier to bypass safety-aligned LLM than single-turn jailbreaks. Existing multi-turn benchmarks are limited in size or rely heavily on templates, which restrict their diversity. To address this gap, we unify a wide range of harmful jailbreak intents, and introduce an active learning pipeline for expanding high-quality multi-turn adversarial prompts, where a generator is iteratively fine-tuned to produce stronger attack candidates, guided by uncertainty-based refinement. Our MultiBreak includes 10,389 multi-turn adversarial prompts, spans 2,665 distinct harmful intents, and covers the most diverse set of topics to date. Empirical evaluation shows that our benchmark achieves up to a 54.0 and 34.6 higher attack success rate (ASR)} than the second-best dataset on DeepSeek-R1-7B and GPT-4.1-mini, respectively. More importantly, safety evaluations suggest that diverse attack categories uncover fine-grained LLM vulnerabilities}, and categories that appear benign under single-turn can exhibit substantially higher adversarial effectiveness in multi-turn scenarios. These findings highlight persistent vulnerabilities of LLMs under realistic adversarial settings and establish MultiBreak as a scalable resource for advancing LLM safety.

0 citations0 reviewsNo code linkedOpen access
preprint2024arXiv

DeepSeek LLM: Scaling Open-Source Language Models with Longtermism

The rapid development of open-source large language models (LLMs) has been truly remarkable. However, the scaling law described in previous literature presents varying conclusions, which casts a dark cloud over scaling LLMs. We delve into the study of scaling laws and present our distinctive findings that facilitate scaling of large scale models in two commonly used open-source configurations, 7B and 67B. Guided by the scaling laws, we introduce DeepSeek LLM, a project dedicated to advancing open-source language models with a long-term perspective. To support the pre-training phase, we have developed a dataset that currently consists of 2 trillion tokens and is continuously expanding. We further conduct supervised fine-tuning (SFT) and Direct Preference Optimization (DPO) on DeepSeek LLM Base models, resulting in the creation of DeepSeek Chat models. Our evaluation results demonstrate that DeepSeek LLM 67B surpasses LLaMA-2 70B on various benchmarks, particularly in the domains of code, mathematics, and reasoning. Furthermore, open-ended evaluations reveal that DeepSeek LLM 67B Chat exhibits superior performance compared to GPT-3.5.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

A Survey of Knowledge-Intensive NLP with Pre-Trained Language Models

With the increasing of model capacity brought by pre-trained language models, there emerges boosting needs for more knowledgeable natural language processing (NLP) models with advanced functionalities including providing and making flexible use of encyclopedic and commonsense knowledge. The mere pre-trained language models, however, lack the capacity of handling such knowledge-intensive NLP tasks alone. To address this challenge, large numbers of pre-trained language models augmented with external knowledge sources are proposed and in rapid development. In this paper, we aim to summarize the current progress of pre-trained language model-based knowledge-enhanced models (PLMKEs) by dissecting their three vital elements: knowledge sources, knowledge-intensive NLP tasks, and knowledge fusion methods. Finally, we present the challenges of PLMKEs based on the discussion regarding the three elements and attempt to provide NLP practitioners with potential directions for further research.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Adversarial Regularization as Stackelberg Game: An Unrolled Optimization Approach

Adversarial regularization has been shown to improve the generalization performance of deep learning models in various natural language processing tasks. Existing works usually formulate the method as a zero-sum game, which is solved by alternating gradient descent/ascent algorithms. Such a formulation treats the adversarial and the defending players equally, which is undesirable because only the defending player contributes to the generalization performance. To address this issue, we propose Stackelberg Adversarial Regularization (SALT), which formulates adversarial regularization as a Stackelberg game. This formulation induces a competition between a leader and a follower, where the follower generates perturbations, and the leader trains the model subject to the perturbations. Different from conventional approaches, in SALT, the leader is in an advantageous position. When the leader moves, it recognizes the strategy of the follower and takes the anticipated follower's outcomes into consideration. Such a leader's advantage enables us to improve the model fitting to the unperturbed data. The leader's strategic information is captured by the Stackelberg gradient, which is obtained using an unrolling algorithm. Our experimental results on a set of machine translation and natural language understanding tasks show that SALT outperforms existing adversarial regularization baselines across all tasks. Our code is available at https://github.com/SimiaoZuo/Stackelberg-Adv.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Analysis of Wall Heat Flux of a Hypersonic Shock Wave / Boundary Layer Interaction with a Novel Decomposition Formula

The generation mechanism of wall heat flux is one of the fundamental problems in supersonic/hypersonic turbulent boundary layers. A novel heat decomposition formula under the curvilinear coordinate was proposed in this paper. The new formula has wider application scope and can be applied in the configurations with grid deformed. The wall heat flux of an interaction between shock wave and the turbulent boundary layer over a compression corner is analyzed by the new formula. The results indicated good performance of the formula in the complex interaction region. The contributions of different energy transport processes were obtained. The contributions by the turbulent fluctuations e.g., Reynolds stresses and turbulent transport of heat flux, were significantly increased, while the processes by the mean profile e.g., molecular stresses and heat conduction, can be neglected. In addition, the pressure work is another contributor to the wall heat flux and the streamwise component works mainly in the shear layer and the reattachment point, while pressure in the wall-normal direction is concentrated in the vicinity of the reattachment point.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

ARCH: Efficient Adversarial Regularized Training with Caching

Adversarial regularization can improve model generalization in many natural language processing tasks. However, conventional approaches are computationally expensive since they need to generate a perturbation for each sample in each epoch. We propose a new adversarial regularization method ARCH (adversarial regularization with caching), where perturbations are generated and cached once every several epochs. As caching all the perturbations imposes memory usage concerns, we adopt a K-nearest neighbors-based strategy to tackle this issue. The strategy only requires caching a small amount of perturbations, without introducing additional training time. We evaluate our proposed method on a set of neural machine translation and natural language understanding tasks. We observe that ARCH significantly eases the computational burden (saves up to 70% of computational time in comparison with conventional approaches). More surprisingly, by reducing the variance of stochastic gradients, ARCH produces a notably better (in most of the tasks) or comparable model generalization. Our code is available at https://github.com/SimiaoZuo/Caching-Adv.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Configuration of the Martian dust rings: Shapes, densities and size-distributions from direct integrations of particle trajectories

It is expected since the early 1970s that tenuous dust rings are formed by grains ejected from the Martian moons Phobos and Deimos by impacts of hypervelocity interplanetary projectiles. In this paper, we perform direct numerical integrations of a large number of dust particles originating from Phobos and Deimos. In the numerical simulations, the most relevant forces acting on dust are included: Martian gravity with spherical harmonics up to 5th degree and 5th order, gravitational perturbations from the Sun, Phobos, and Deimos, solar radiation pressure, as well as the Poynting-Robertson drag. In order to obtain the ring configuration, simulation results of various grain sizes ranging from submicron to 100 microns are averaged over a specified initial mass distribution of ejecta. We find that for the Phobos ring grains smaller than about 2 microns are dominant; while the Deimos ring is dominated by dust in the size range of about 5-20 microns. The asymmetries, number densities and geometrical optical depths of the rings are quantified from simulations. The results are compared with the upper limits of the optical depth inferred from Hubble observations. We compare to previous work and discuss the uncertainties of the models.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Distribution of dust ejected from the lunar surface into the Earth-Moon system

Aims. An asymmetric dust cloud was detected around the Moon by the Lunar Dust Experiment on board the Lunar Atmosphere and Dust Environment Explorer mission. We investigate the dynamics of the grains that escape the Moon and their configuration in the Earth-Moon system. Methods. We use a plausible initial ejecta distribution and mass production rate for the ejected dust. Various forces, including the solar radiation pressure and the gravity of the Moon, Earth, and Sun, are considered in the dynamical model, and direct numerical integrations of trajectories of dust particles are performed. The final states, the average life spans, and the fraction of retrograde grains as functions of particle size are computed. The number density distribution in the Earth-Moon system is obtained through long-term simulations. Results. The average life spans depend on the size of dust particles and show a rapid increase in the size range between $1\, \mathrm{μm}$ and $10\, \mathrm{μm}$. About ${3.6\times10^{-3}\,\mathrm{kg/s}}$ ($\sim2\%$) particles ejected from the lunar surface escape the gravity of the Moon, and they form an asymmetric torus between the Earth and the Moon in the range $[10\,R_\mathrm{E},50\,R_\mathrm{E}]$, which is offset toward the direction of the Sun. A considerable number of retrograde particles occur in the Earth-Moon system.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Human Parity on CommonsenseQA: Augmenting Self-Attention with External Attention

Most of today's AI systems focus on using self-attention mechanisms and transformer architectures on large amounts of diverse data to achieve impressive performance gains. In this paper, we propose to augment the transformer architecture with an external attention mechanism to bring external knowledge and context to bear. By integrating external information into the prediction process, we hope to reduce the need for ever-larger models and increase the democratization of AI systems. We find that the proposed external attention mechanism can significantly improve the performance of existing AI systems, allowing practitioners to easily customize foundation AI models to many diverse downstream applications. In particular, we focus on the task of Commonsense Reasoning, demonstrating that the proposed external attention mechanism can augment existing transformer models and significantly improve the model's reasoning capabilities. The proposed system, Knowledgeable External Attention for commonsense Reasoning (KEAR), reaches human parity on the open CommonsenseQA research benchmark with an accuracy of 89.4\% in comparison to the human accuracy of 88.9\%.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Knowledge-Rich Self-Supervision for Biomedical Entity Linking

Entity linking faces significant challenges such as prolific variations and prevalent ambiguities, especially in high-value domains with myriad entities. Standard classification approaches suffer from the annotation bottleneck and cannot effectively handle unseen entities. Zero-shot entity linking has emerged as a promising direction for generalizing to new entities, but it still requires example gold entity mentions during training and canonical descriptions for all entities, both of which are rarely available outside of Wikipedia. In this paper, we explore Knowledge-RIch Self-Supervision ($\tt KRISS$) for biomedical entity linking, by leveraging readily available domain knowledge. In training, it generates self-supervised mention examples on unlabeled text using a domain ontology and trains a contextual encoder using contrastive learning. For inference, it samples self-supervised mentions as prototypes for each entity and conducts linking by mapping the test mention to the most similar prototype. Our approach can easily incorporate entity descriptions and gold mention labels if available. We conducted extensive experiments on seven standard datasets spanning biomedical literature and clinical notes. Without using any labeled information, our method produces $\tt KRISSBERT$, a universal entity linker for four million UMLS entities that attains new state of the art, outperforming prior self-supervised methods by as much as 20 absolute points in accuracy.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

LiST: Lite Prompted Self-training Makes Parameter-Efficient Few-shot Learners

We present a new method LiST is short for Lite Prompted Self-Training for parameter-efficient fine-tuning of large pre-trained language models (PLMs) for few-shot learning. LiST improves over recent methods that adopt prompt-based fine-tuning (FN) using two key techniques. The first is the use of self-training to leverage large amounts of unlabeled data for prompt-based FN in few-shot settings. We use self-training in conjunction with meta-learning for re-weighting noisy pseudo-prompt labels. Self-training is expensive as it requires updating all the model parameters repetitively. Therefore, we use a second technique for light-weight fine-tuning where we introduce a small number of task-specific parameters that are fine-tuned during self-training while keeping the PLM encoder frozen. Our experiments show that LiST can effectively leverage unlabeled data to improve the model performance for few-shot learning. Additionally, the fine-tuning is efficient as it only updates a small percentage of parameters and the overall model footprint is reduced since several tasks can share a common PLM encoder as backbone. A comprehensive study on six NLU tasks demonstrate LiST to improve by 35% over classic fine-tuning and 6% over prompt-based FN with 96% reduction in number of trainable parameters when fine-tuned with no more than 30 labeled examples from each task. With only 14M tunable parameters, LiST outperforms GPT-3 in-context learning by 33% on few-shot NLU tasks.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

METRO: Efficient Denoising Pretraining of Large Scale Autoencoding Language Models with Model Generated Signals

We present an efficient method of pretraining large-scale autoencoding language models using training signals generated by an auxiliary model. Originated in ELECTRA, this training strategy has demonstrated sample-efficiency to pretrain models at the scale of hundreds of millions of parameters. In this work, we conduct a comprehensive empirical study, and propose a recipe, namely "Model generated dEnoising TRaining Objective" (METRO), which incorporates some of the best modeling techniques developed recently to speed up, stabilize, and enhance pretrained language models without compromising model effectiveness. The resultant models, METRO-LM, consisting of up to 5.4 billion parameters, achieve new state-of-the-art on the GLUE, SuperGLUE, and SQuAD benchmarks. More importantly, METRO-LM are efficient in that they often outperform previous large models with significantly smaller model sizes and lower pretraining cost.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

No Parameters Left Behind: Sensitivity Guided Adaptive Learning Rate for Training Large Transformer Models

Recent research has shown the existence of significant redundancy in large Transformer models. One can prune the redundant parameters without significantly sacrificing the generalization performance. However, we question whether the redundant parameters could have contributed more if they were properly trained. To answer this question, we propose a novel training strategy that encourages all parameters to be trained sufficiently. Specifically, we adaptively adjust the learning rate for each parameter according to its sensitivity, a robust gradient-based measure reflecting this parameter's contribution to the model performance. A parameter with low sensitivity is redundant, and we improve its fitting by increasing its learning rate. In contrast, a parameter with high sensitivity is well-trained, and we regularize it by decreasing its learning rate to prevent further overfitting. We conduct extensive experiments on natural language understanding, neural machine translation, and image classification to demonstrate the effectiveness of the proposed schedule. Analysis shows that the proposed schedule indeed reduces the redundancy and improves generalization performance.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Open Domain Question Answering with A Unified Knowledge Interface

The retriever-reader framework is popular for open-domain question answering (ODQA) due to its ability to use explicit knowledge. Although prior work has sought to increase the knowledge coverage by incorporating structured knowledge beyond text, accessing heterogeneous knowledge sources through a unified interface remains an open question. While data-to-text generation has the potential to serve as a universal interface for data and text, its feasibility for downstream tasks remains largely unknown. In this work, we bridge this gap and use the data-to-text method as a means for encoding structured knowledge for ODQA. Specifically, we propose a verbalizer-retriever-reader framework for ODQA over data and text where verbalized tables from Wikipedia and graphs from Wikidata are used as augmented knowledge sources. We show that our Unified Data and Text QA, UDT-QA, can effectively benefit from the expanded knowledge index, leading to large gains over text-only baselines. Notably, our approach sets the single-model state-of-the-art on Natural Questions. Furthermore, our analyses indicate that verbalized knowledge is preferred for answer reasoning for both adapted and hot-swap settings.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Review of research on lunar dust dynamics

Lunar dust particles are generated by hypervelocity impacts of interplanetary micron-meteoroids onto the surface of the Moon, which seriously threatens the security of explorations. Studying the lunar dust dynamics helps to understand the origin and migration mechanism of lunar dust, and to provide the theoretical guidelines for the orbital design of lunar space missions. This paper reviews previous research on the lunar dust dynamics, including the interplanetary impactor environment at the Earth-Moon system, the mass production rate, the initial mass, speed and ejecta angle distributions, the related space exploration missions, the dynamical model and spatial distribution of dust particles originating from the lunar surface in the whole Earth-Moon system.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Spectral and Energy Efficiency of DCO-OFDM in Visible Light Communication Systems with Finite-Alphabet Inputs

The bound of the information transmission rate of direct current biased optical orthogonal frequency division multiplexing (DCO-OFDM) for visible light communication (VLC) with finite-alphabet inputs is yet unknown, where the corresponding spectral efficiency (SE) and energy efficiency (EE) stems out as the open research problems. In this paper, we derive the exact achievable rate of {the} DCO-OFDM system with finite-alphabet inputs for the first time. Furthermore, we investigate SE maximization problems of {the} DCO-OFDM system subject to both electrical and optical power constraints. By exploiting the relationship between the mutual information and the minimum mean-squared error, we propose a multi-level mercury-water-filling power allocation scheme to achieve the maximum SE. Moreover, the EE maximization problems of {the} DCO-OFDM system are studied, and the Dinkelbach-type power allocation scheme is developed for the maximum EE. Numerical results verify the effectiveness of the proposed theories and power allocation schemes.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Taming Sparsely Activated Transformer with Stochastic Experts

Sparsely activated models (SAMs), such as Mixture-of-Experts (MoE), can easily scale to have outrageously large amounts of parameters without significant increase in computational cost. However, SAMs are reported to be parameter inefficient such that larger models do not always lead to better performance. While most on-going research focuses on improving SAMs models by exploring methods of routing inputs to experts, our analysis reveals that such research might not lead to the solution we expect, i.e., the commonly-used routing methods based on gating mechanisms do not work better than randomly routing inputs to experts. In this paper, we propose a new expert-based model, THOR (Transformer witH StOchastic ExpeRts). Unlike classic expert-based models, such as the Switch Transformer, experts in THOR are randomly activated for each input during training and inference. THOR models are trained using a consistency regularized loss, where experts learn not only from training data but also from other experts as teachers, such that all the experts make consistent predictions. We validate the effectiveness of THOR on machine translation tasks. Results show that THOR models are more parameter efficient in that they significantly outperform the Transformer and MoE models across various settings. For example, in multilingual translation, THOR outperforms the Switch Transformer by 2 BLEU scores, and obtains the same BLEU score as that of a state-of-the-art MoE model that is 18 times larger. Our code is publicly available at: https://github.com/microsoft/Stochastic-Mixture-of-Experts.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Tensor Programs V: Tuning Large Neural Networks via Zero-Shot Hyperparameter Transfer

Hyperparameter (HP) tuning in deep learning is an expensive process, prohibitively so for neural networks (NNs) with billions of parameters. We show that, in the recently discovered Maximal Update Parametrization (muP), many optimal HPs remain stable even as model size changes. This leads to a new HP tuning paradigm we call muTransfer: parametrize the target model in muP, tune the HP indirectly on a smaller model, and zero-shot transfer them to the full-sized model, i.e., without directly tuning the latter at all. We verify muTransfer on Transformer and ResNet. For example, 1) by transferring pretraining HPs from a model of 13M parameters, we outperform published numbers of BERT-large (350M parameters), with a total tuning cost equivalent to pretraining BERT-large once; 2) by transferring from 40M parameters, we outperform published numbers of the 6.7B GPT-3 model, with tuning cost only 7% of total pretraining cost. A Pytorch implementation of our technique can be found at github.com/microsoft/mup and installable via `pip install mup`.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

A discontinuous Galerkin method based on a hierarchical orthogonal basis for Lagrangian hydrodynamics on curvilinear grids

We present a new high-order accurate Lagrangian discontinuous Galerkin (DG) hydrodynamic method to simulate material dynamics (for e.g., gasses, fluids, and solids) with up to fourth-order accuracy on cubic meshes. The variables, such as specific volume, velocity, specific total energy, and deformation gradient fields within a cell, are represented with a polynomial constructed from a novel hierarchical orthogonal basis about the center of mass, which decouples the moments of the solution because the mass matrix is diagonal. The discontinuity in the polynomials at the cell boundary is addressed by solving a multi-directional Riemann problem at the vertices of the cell and a 1D Riemann problem at additional non-vertex quadrature points along the edges so that the surface integral is exact for the polynomial order. The uniqueness lies in that the vertices of the curvilinear grid work as the quadrature points for the surface integral of DG methods. To ensure robust mesh motion, the pressure for the Riemann problem accounts for the difference between the density variation over the cell and a density field from subcell mesh stabilization (SMS). The accuracy and robustness of the new high-order accurate Lagrangian DG hydrodynamic method is demonstrated by simulating a diverse suite of challenging test problems covering gas and solid dynamic problems on curved grids.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Data completion algorithms and their applications in inverse acoustic scattering with limited-aperture backscattering data

We introduce two data completion algorithms for the limited-aperture problems in inverse acoustic scattering. Both completion algorithms are independent of the topological and physical properties of the unknown scatterers. The main idea is to relate the limited-aperture data to the full-aperture data via the prolate matrix. The data completion algorithms are simple and fast since only the approximate inversion of the prolate matrix is involved. We then combine the data completion algorithms with imaging methods such as factorization method and direct sampling method for the object reconstructions. A variety of numerical examples are presented to illustrate the effectiveness and robustness of the proposed algorithms.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

DepOwl: Detecting Dependency Bugs to Prevent Compatibility Failures

Applications depend on libraries to avoid reinventing the wheel. Libraries may have incompatible changes during evolving. As a result, applications will suffer from compatibility failures. There has been much research on addressing detecting incompatible changes in libraries, or helping applications co-evolve with the libraries. The existing solution helps the latest application version work well against the latest library version as an afterthought. However, end users have already been suffering from the failures and have to wait for new versions. In this paper, we propose DepOwl, a practical tool helping users prevent compatibility failures. The key idea is to avoid using incompatible versions from the very beginning. We evaluated DepOwl on 38 known compatibility failures from StackOverflow, and DepOwl can prevent 32 of them. We also evaluated DepOwl using the software repository shipped with Ubuntu-19.10. DepOwl detected 77 unknown dependency bugs, which may lead to compatibility failures.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Adversarial Training for Commonsense Inference

We propose an AdversariaL training algorithm for commonsense InferenCE (ALICE). We apply small perturbations to word embeddings and minimize the resultant adversarial risk to regularize the model. We exploit a novel combination of two different approaches to estimate these perturbations: 1) using the true label and 2) using the model prediction. Without relying on any human-crafted features, knowledge bases, or additional datasets other than the target datasets, our model boosts the fine-tuning performance of RoBERTa, achieving competitive results on multiple reading comprehension datasets that require commonsense inference.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Adversarial Training for Large Neural Language Models

Generalization and robustness are both key desiderata for designing machine learning methods. Adversarial training can enhance robustness, but past work often finds it hurts generalization. In natural language processing (NLP), pre-training large neural language models such as BERT have demonstrated impressive gain in generalization for a variety of tasks, with further improvement from adversarial fine-tuning. However, these models are still vulnerable to adversarial attacks. In this paper, we show that adversarial pre-training can improve both generalization and robustness. We propose a general algorithm ALUM (Adversarial training for large neural LangUage Models), which regularizes the training objective by applying perturbations in the embedding space that maximizes the adversarial loss. We present the first comprehensive study of adversarial training in all stages, including pre-training from scratch, continual pre-training on a well-trained model, and task-specific fine-tuning. ALUM obtains substantial gains over BERT on a wide range of NLP tasks, in both regular and adversarial scenarios. Even for models that have been well trained on extremely large text corpora, such as RoBERTa, ALUM can still produce significant gains from continual pre-training, whereas conventional non-adversarial methods can not. ALUM can be further combined with task-specific fine-tuning to attain additional gains. The ALUM code is publicly available at https://github.com/namisan/mt-dnn.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Identification of point like objects with multi-frequency sparse data

The inverse acoustic scattering of point objects using multi-frequency sparse measurements are studied. The objects may be a sum of point sources or point like scatterers. We show that the locations and scattering strengths of the point objects can be uniquely identified by the multi-frequency near or far fields taken at sparse sensors. Based on the uniqueness analysis, some direct methods have also been proposed for reconstructing the locations and determining the scattering strengths. The numerical examples are conducted to show the validity and robustness of the proposed numerical methods.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

MLFcGAN: Multi-level Feature Fusion based Conditional GAN for Underwater Image Color Correction

Color correction for underwater images has received increasing interests, due to its critical role in facilitating available mature vision algorithms for underwater scenarios. Inspired by the stunning success of deep convolutional neural networks (DCNNs) techniques in many vision tasks, especially the strength in extracting features in multiple scales, we propose a deep multi-scale feature fusion net based on the conditional generative adversarial network (GAN) for underwater image color correction. In our network, multi-scale features are extracted first, followed by augmenting local features on each scale with global features. This design was verified to facilitate more effective and faster network learning, resulting in better performance in both color correction and detail preservation. We conducted extensive experiments and compared with the state-of-the-art approaches quantitatively and qualitatively, showing that our method achieves significant improvements.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Multi-Modality Cascaded Fusion Technology for Autonomous Driving

Multi-modality fusion is the guarantee of the stability of autonomous driving systems. In this paper, we propose a general multi-modality cascaded fusion framework, exploiting the advantages of decision-level and feature-level fusion, utilizing target position, size, velocity, appearance and confidence to achieve accurate fusion results. In the fusion process, dynamic coordinate alignment(DCA) is conducted to reduce the error between sensors from different modalities. In addition, the calculation of affinity matrix is the core module of sensor fusion, we propose an affinity loss that improves the performance of deep affinity network(DAN). Last, the proposed step-by-step cascaded fusion framework is more interpretable and flexible compared to the end-toend fusion methods. Extensive experiments on Nuscenes [2] dataset show that our approach achieves the state-of-theart performance.dataset show that our approach achieves the state-of-the-art performance.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

The Microsoft Toolkit of Multi-Task Deep Neural Networks for Natural Language Understanding

We present MT-DNN, an open-source natural language understanding (NLU) toolkit that makes it easy for researchers and developers to train customized deep learning models. Built upon PyTorch and Transformers, MT-DNN is designed to facilitate rapid customization for a broad spectrum of NLU tasks, using a variety of objectives (classification, regression, structured prediction) and text encoders (e.g., RNNs, BERT, RoBERTa, UniLM). A unique feature of MT-DNN is its built-in support for robust and transferable learning using the adversarial multi-task learning paradigm. To enable efficient production deployment, MT-DNN supports multi-task knowledge distillation, which can substantially compress a deep neural model without significant performance drop. We demonstrate the effectiveness of MT-DNN on a wide range of NLU applications across general and biomedical domains. The software and pre-trained models will be publicly available at https://github.com/namisan/mt-dnn.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

UniLMv2: Pseudo-Masked Language Models for Unified Language Model Pre-Training

We propose to pre-train a unified language model for both autoencoding and partially autoregressive language modeling tasks using a novel training procedure, referred to as a pseudo-masked language model (PMLM). Given an input text with masked tokens, we rely on conventional masks to learn inter-relations between corrupted tokens and context via autoencoding, and pseudo masks to learn intra-relations between masked spans via partially autoregressive modeling. With well-designed position embeddings and self-attention masks, the context encodings are reused to avoid redundant computation. Moreover, conventional masks used for autoencoding provide global masking information, so that all the position embeddings are accessible in partially autoregressive language modeling. In addition, the two tasks pre-train a unified language model as a bidirectional encoder and a sequence-to-sequence decoder, respectively. Our experiments show that the unified language models pre-trained using PMLM achieve new state-of-the-art results on a wide range of natural language understanding and generation tasks across several widely used benchmarks.

0 citations0 reviewsNo code linkedOpen access
preprint2018arXiv

Dust in the Jupiter system outside the rings

Jupiter is one of the major targets for planetary exploration, and dust in the Jovian system is of great interest to researchers in the field of planetary science. In this paper, we review the five dust populations outside the ring system: grains in the region of the Galilean moons, potential dust from plumes on Europa, Jovian stream particles, particles in the outer region of the Jovian system ejected from the irregular satellites, and dust in the region of the Trojan asteroids. The physical environment for the dust dynamics is described, including the gravity, the magnetic field and the plasma environment. For each population, the dust sources are described, and the relevant perturbation forces are discussed. Observations and results from modeling are reviewed, and the distributions of the individual dust populations are shown. The understanding of the Jovian dust environment allows to assess the dust hazard to spacecraft, and to characterize the material exchange between the Jovian moons, their surface properties and distribution of non-icy constituents.

0 citations0 reviewsNo code linkedOpen access
preprint2018arXiv

Inverse elastic scattering problems with phaseless far field data

This paper is concerned with uniqueness, phase retrieval and shape reconstruction methods for inverse elastic scattering problems with phaseless far field data. Systematically, we study two basic models, i.e., inverse scattering of plane waves by rigid bodies and inverse scattering of sources with compact support. For both models, we show that the location of the objects can not be uniquely recovered by the data. To solve this problem, we consider simultaneously the incident point sources with one fixed source point and at most three scattering strengths. We then establish some uniqueness results for source scattering problem with multi-frequency phaseless far field data. Furthermore, a fast and stable phase retrieval approach is proposed based on a simple geometric result which provides a stable reconstruction of a point in the plane from three distances to given points. Difficulties arise for inverse scattering by rigid bodies due to the additional unknown far field pattern of the point sources. To overcome this difficulty, we introduce an artificial rigid body into the system and show that the underlying rigid bodies can be uniquely determined by the corresponding phaseless far field data at a fixed frequency. Noting that the far field pattern of the scattered field corresponding to point sources is very small if the source point is far away from the scatterers, we propose an appropriate phase retrieval method for obstacle scattering problems, without using the artificial rigid body. Finally, we propose several sampling methods for shape reconstruction with phaseless far field data. Extended numerical examples in two dimensions are conducted with noisy data, and the results further verify the effectiveness and robustness of the proposed phase retrieval techniques and sampling methods.

0 citations0 reviewsNo code linkedOpen access
preprint2018arXiv

On a novel inverse scattering scheme using resonant modes with enhanced imaging resolution

We develop a novel wave imaging scheme for reconstructing the shape of an inhomogeneous scatterer and we consider the inverse acoustic obstacle scattering problem as a prototype model for our study. There exists a wealth of reconstruction methods for the inverse obstacle scattering problem and many of them intentionally avoid the interior resonant modes. Indeed, the occurrence of the interior resonance may cause the failure of the corresponding reconstruction. However, based on the observation that the interior resonant modes actually carry the geometrical information of the underlying obstacle, we propose an inverse scattering scheme of using those resonant modes for the reconstruction. To that end, we first develop a numerical procedure in determining the interior eigenvalues associated with an unknown obstacle from its far-field data based on the validity of the factorization method. Then we propose two efficient optimization methods in further determining the corresponding eigenfunctions. Using the afore-determined interior resonant modes, we show that the shape of the underlying obstacle can be effectively recovered. Moreover, the reconstruction yields enhanced imaging resolution, especially for the concave part of the obstacle. We provide rigorous theoretical justifications for the proposed method. Numerical examples in 2D and 3D verify the theoretically predicted effectiveness and efficiency of the method.

0 citations0 reviewsNo code linkedOpen access