arXiv subjects

Jin Chen

Publications and source records attributed to Jin Chen.

At least 55 records · Page 3Linked to original sources

arXiv2026

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.

cs.CL↗

arXiv2025

WholeBodyVLA: Towards Unified Latent VLA for Whole-Body Loco-Manipulation Control

Humanoid robots require precise locomotion and dexterous manipulation to perform challenging loco-manipulation tasks. Yet existing approaches, modular or end-to-end, are deficient in manipulation-aware locomotion. This confines the robot to a limited workspace, preventing it from performing large-space loco-manipulation. We attribute this to: (1) the challenge of acquiring loco-manipulation knowledge due to the scarcity of humanoid teleoperation data, and (2) the difficulty of faithfully and reliably executing locomotion commands, stemming from the limited precision and stability of existing RL controllers. To acquire richer loco-manipulation knowledge, we propose a unified latent learning framework that enables Vision-Language-Action (VLA) system to learn from low-cost action-free egocentric videos. Moreover, an efficient human data collection pipeline is devised to augment the dataset and scale the benefits. To execute the desired locomotion commands more precisely, we present a loco-manipulation-oriented (LMO) RL policy specifically tailored for accurate and stable core loco-manipulation movements, such as advancing, turning, and squatting. Building on these components, we introduce WholeBodyVLA, a unified framework for humanoid loco-manipulation. To the best of our knowledge, WholeBodyVLA is one of its kind enabling large-space humanoid loco-manipulation. It is verified via comprehensive experiments on the AgiBot X2 humanoid, outperforming prior baseline by 21.3%. It also demonstrates strong generalization and high extensibility across a broad range of tasks.

cs.RO↗

arXiv2025

Adaptive Evidential Learning for Temporal-Semantic Robustness in Moment Retrieval

In the domain of moment retrieval, accurately identifying temporal segments within videos based on natural language queries remains challenging. Traditional methods often employ pre-trained models that struggle with fine-grained information and deterministic reasoning, leading to difficulties in aligning with complex or ambiguous moments. To overcome these limitations, we explore Deep Evidential Regression (DER) to construct a vanilla Evidential baseline. However, this approach encounters two major issues: the inability to effectively handle modality imbalance and the structural differences in DER's heuristic uncertainty regularizer, which adversely affect uncertainty estimation. This misalignment results in high uncertainty being incorrectly associated with accurate samples rather than challenging ones. Our observations indicate that existing methods lack the adaptability required for complex video scenarios. In response, we propose Debiased Evidential Learning for Moment Retrieval (DEMR), a novel framework that incorporates a Reflective Flipped Fusion (RFF) block for cross-modal alignment and a query reconstruction task to enhance text sensitivity, thereby reducing bias in uncertainty estimation. Additionally, we introduce a Geom-regularizer to refine uncertainty predictions, enabling adaptive alignment with difficult moments and improving retrieval accuracy. Extensive testing on standard datasets and debiased datasets ActivityNet-CD and Charades-CD demonstrates significant enhancements in effectiveness, robustness, and interpretability, positioning our approach as a promising solution for temporal-semantic robustness in moment retrieval. The code is publicly available at https://github.com/KaijingOfficial/DEMR.

cs.CV↗

arXiv2025

Fermionic Non-invertible Symmetry Behind Supersymmetric ADE Solitons

The non-perturbative constraints imposed by intrinsic fermionic non-invertible symmetries in 1+1 dimensional gapped systems remain largely unexplored. In this letter, we propose the superstrip algebra as a unified framework to catalog the categorical symmetry data in a massive fermionic model. The algebra and its representations explicitly encode the vacuum structure, soliton degeneracies, and their quantum numbers. As a demonstration, we apply this framework to the $\mathcal N=2$ minimal models with their least relevant deformation. We show that this specific deformation alone preserves a non-invertible superfusion category, a fermionic variant of $\text{SU}(2)_k$ known to underlie the $ADE$ classification of critical theories. Its superstrip algebra then accounts for the origin of the resulting $ADE$-type soliton spectrum and their fractional fermion number. Although our primary examples are supersymmetric and integrable, our framework itself relies on neither property, providing a new powerful tool for studying a broad class of strongly-coupled fermionic systems.

hep-th↗

arXiv2025

Virtual Width Networks

We introduce Virtual Width Networks (VWN), a framework that delivers the benefits of wider representations without incurring the quadratic cost of increasing the hidden size. VWN decouples representational width from backbone width, expanding the embedding space while keeping backbone compute nearly constant. In our large-scale experiment, an 8-times expansion accelerates optimization by over 2 times for next-token and 3 times for next-2-token prediction. The advantage amplifies over training as both the loss gap grows and the convergence-speedup ratio increases, showing that VWN is not only token-efficient but also increasingly effective with scale. Moreover, we identify an approximately log-linear scaling relation between virtual width and loss reduction, offering an initial empirical basis and motivation for exploring virtual-width scaling as a new dimension of large-model efficiency.

cs.LG↗

arXiv2025

Poisson Flow Consistency Training

The Poisson Flow Consistency Model (PFCM) is a consistency-style model based on the robust Poisson Flow Generative Model++ (PFGM++) which has achieved success in unconditional image generation and CT image denoising. Yet the PFCM can only be trained in distillation which limits the potential of the PFCM in many data modalities. The objective of this research was to create a method to train the PFCM in isolation called Poisson Flow Consistency Training (PFCT). The perturbation kernel was leveraged to remove the pretrained PFGM++, and the sinusoidal discretization schedule and Beta noise distribution were introduced in order to facilitate adaptability and improve sample quality. The model was applied to the task of low dose computed tomography image denoising and improved the low dose image in terms of LPIPS and SSIM. It also displayed similar denoising effectiveness as models like the Consistency Model. PFCT is established as a valid method of training the PFCM from its effectiveness in denoising CT images, showing potential with competitive results to other generative models. Further study is needed in the precise optimization of PFCT and in its applicability to other generative modeling tasks. The framework of PFCT creates more flexibility for the ways in which a PFCM can be created and can be applied to the field of generative modeling.

cs.CV↗

arXiv2025

Beyond the Seen: Bounded Distribution Estimation for Open-Vocabulary Learning

Open-vocabulary learning requires modeling the data distribution in open environments, which consists of both seen-class and unseen-class data. Existing methods estimate the distribution in open environments using seen-class data, where the absence of unseen classes makes the estimation error inherently unidentifiable. Intuitively, learning beyond the seen classes is crucial for distribution estimation to bound the estimation error. We theoretically demonstrate that the distribution can be effectively estimated by generating unseen-class data, through which the estimation error is upper-bounded. Building on this theoretical insight, we propose a novel open-vocabulary learning method, which generates unseen-class data for estimating the distribution in open environments. The method consists of a class-domain-wise data generation pipeline and a distribution alignment algorithm. The data generation pipeline generates unseen-class data under the guidance of a hierarchical semantic tree and domain information inferred from the seen-class data, facilitating accurate distribution estimation. With the generated data, the distribution alignment algorithm estimates and maximizes the posterior probability to enhance generalization in open-vocabulary learning. Extensive experiments on $11$ datasets demonstrate that our method outperforms baseline approaches by up to $14\%$, highlighting its effectiveness and superiority.

cs.CV↗

arXiv2025

Dispersion Formation Control: from Geometry to Distribution

We introduce and develop the concept of dispersion formation control, bridging a gap between shape-assembly studies in physics and biology and formation control theory. In current formation control studies, the control objectives typically focus on achieving desired local geometric properties, such as inter-agent distances, bearings, or relative positions. In contrast, our dispersion formation control approach enables agents to directly regulate the dispersion of their spatial distribution, a global variable associated with a covariance matrix. Specifically, we introduce the notion of covariance similarity to define the target spatial dispersion of agents. Building on this framework, we propose two control strategies: a centralized approach to illustrate the key ideas, and a distributed approach that enables agents to control the global dispersion but using only local information. Our stability analysis demonstrates that both strategies ensure exponential convergence of the agents' distribution to the desired dispersion. Notably, controlling a global variable rather than multiple local ones enhances the resiliency of the system, particularly against malfunctioning agents. Simulations validate the effectiveness of the proposed dispersion formation control.

eess.SY↗

arXiv2025

Comparative analysis of plasmon modes in layered Lindhard metals and strange metals

The enigmatic strange metal remains one of the central unsolved problems of 21st century science. Understanding this phase of matter requires knowledge of the momentum- and energy-resolved dynamic charge susceptibility, $χ(q,ω)$, especially at finite momentum. Inelastic electron scattering (EELS), performed in either transmission (T-EELS) or reflection (R-EELS) geometries, is a powerful probe of $χ(q,ω)$. For the prototypical strange metal Bi$_2$Sr$_2$CaCu$_2$O$_{8+x}$, T-EELS, R-EELS, and infrared (IR) spectroscopy agree at $q \sim 0$, all revealing a highly damped plasmon near 1 eV. At larger $q$, however, EELS results show unresolved discrepancies. Since IR data are highly reproducible, it is advantageous to use IR data to calculate what the expected EELS response should be at modest $q$. Building on prior R-EELS work [J. Chen \textit{et al.}, Phys. Rev. B. \textbf{109}, 045108 (2024)], we extend this approach to T-EELS for finite stacks of metallic layers, comparing a "textbook" Lindhard metal to a strange metal. In the Lindhard case, the low-$q$ response is dominated by long-lived, standing wave plasmon modes arising from interlayer Coulomb coupling, with in-plane dispersions that resemble the well-known Fetter modes of layered metals. This behavior depends only on the geometry and the long-ranged nature of the Coulomb interaction, and is largely insensitive to layer details. At larger $q$, the response reflects the microscopic properties of individual layers. For the strange metal, calculations based on IR data predict a highly damped plasmon with weak dispersion and no distinct surface mode. While our results match IR and R-EELS at low $q$, they do not reproduce any published EELS spectra at large $q$, highlighting unresolved discrepancies that demand further experimental investigation.

cond-mat.str-el↗

arXiv2025

Simulate, Refocus and Ensemble: An Attention-Refocusing Scheme for Domain Generalization

Domain generalization (DG) aims to learn a model from source domains and apply it to unseen target domains with out-of-distribution data. Owing to CLIP's strong ability to encode semantic concepts, it has attracted increasing interest in domain generalization. However, CLIP often struggles to focus on task-relevant regions across domains, i.e., domain-invariant regions, resulting in suboptimal performance on unseen target domains. To address this challenge, we propose an attention-refocusing scheme, called Simulate, Refocus and Ensemble (SRE), which learns to reduce the domain shift by aligning the attention maps in CLIP via attention refocusing. SRE first simulates domain shifts by performing augmentation on the source data to generate simulated target domains. SRE then learns to reduce the domain shifts by refocusing the attention in CLIP between the source and simulated target domains. Finally, SRE utilizes ensemble learning to enhance the ability to capture domain-invariant attention maps between the source data and the simulated target data. Extensive experimental results on several datasets demonstrate that SRE generally achieves better results than state-of-the-art methods. The code is available at: https://github.com/bitPrincy/SRE-DG.

cs.CV↗

arXiv2025

Detection of Breast Cancer Lumpectomy Margin with SAM-incorporated Forward-Forward Contrastive Learning

Complete removal of cancer tumors with a negative specimen margin during lumpectomy is essential in reducing breast cancer recurrence. However, 2D specimen radiography (SR), the current method used to assess intraoperative specimen margin status, has limited accuracy, resulting in nearly a quarter of patients requiring additional surgery. To address this, we propose a novel deep learning framework combining the Segment Anything Model (SAM) with Forward-Forward Contrastive Learning (FFCL), a pre-training strategy leveraging both local and global contrastive learning for patch-level classification of SR images. After annotating SR images with regions of known maligancy, non-malignant tissue, and pathology-confirmed margins, we pre-train a ResNet-18 backbone with FFCL to classify margin status, then reconstruct coarse binary masks to prompt SAM for refined tumor margin segmentation. Our approach achieved an AUC of 0.8455 for margin classification and segmented margins with a 27.4% improvement in Dice similarity over baseline models, while reducing inference time to 47 milliseconds per image. These results demonstrate that FFCL-SAM significantly enhances both the speed and accuracy of intraoperative margin assessment, with strong potential to reduce re-excision rates and improve surgical outcomes in breast cancer treatment. Our code is available at https://github.com/tbwa233/FFCL-SAM/.

cs.CV↗

arXiv2025

NDCG-Consistent Softmax Approximation with Accelerated Convergence

Ranking tasks constitute fundamental components of extreme similarity learning frameworks, where extremely large corpora of objects are modeled through relative similarity relationships adhering to predefined ordinal structures. Among various ranking surrogates, Softmax (SM) Loss has been widely adopted due to its natural capability to handle listwise ranking via global negative comparisons, along with its flexibility across diverse application scenarios. However, despite its effectiveness, SM Loss often suffers from significant computational overhead and scalability limitations when applied to large-scale object spaces. To address this challenge, we propose novel loss formulations that align directly with ranking metrics: the Ranking-Generalizable \textbf{squared} (RG$^2$) Loss and the Ranking-Generalizable interactive (RG$^\times$) Loss, both derived through Taylor expansions of the SM Loss. Notably, RG$^2$ reveals the intrinsic mechanisms underlying weighted squared losses (WSL) in ranking methods and uncovers fundamental connections between sampling-based and non-sampling-based loss paradigms. Furthermore, we integrate the proposed RG losses with the highly efficient Alternating Least Squares (ALS) optimization method, providing both generalization guarantees and convergence rate analyses. Empirical evaluations on real-world datasets demonstrate that our approach achieves comparable or superior ranking performance relative to SM Loss, while significantly accelerating convergence. This framework offers the similarity learning community both theoretical insights and practically efficient tools, with methodologies applicable to a broad range of tasks where balancing ranking quality and computational efficiency is essential.

cs.LG↗

arXiv2025

VUDG: A Dataset for Video Understanding Domain Generalization

Video understanding has made remarkable progress in recent years, largely driven by advances in deep models and the availability of large-scale annotated datasets. However, existing works typically ignore the inherent domain shifts encountered in real-world video applications, leaving domain generalization (DG) in video understanding underexplored. Hence, we propose Video Understanding Domain Generalization (VUDG), a novel dataset designed specifically for evaluating the DG performance in video understanding. VUDG contains videos from 11 distinct domains that cover three types of domain shifts, and maintains semantic similarity across different domains to ensure fair and meaningful evaluation. We propose a multi-expert progressive annotation framework to annotate each video with both multiple-choice and open-ended question-answer pairs. Extensive experiments on 9 representative large video-language models (LVLMs) and several traditional video question answering methods show that most models (including state-of-the-art LVLMs) suffer performance degradation under domain shifts. These results highlight the challenges posed by VUDG and the difference in the robustness of current models to data distribution shifts. We believe VUDG provides a valuable resource for prompting future research in domain generalization video understanding.

cs.CV↗

arXiv2025

Uncertainty-Guided Self-Questioning and Answering for Video-Language Alignment

The development of multi-modal models has been rapidly advancing, with some demonstrating remarkable capabilities. However, annotating video-text pairs remains expensive and insufficient. Take video question answering (VideoQA) tasks as an example, human annotated questions and answers often cover only part of the video, since the corresponding text is often short and monotonous, leading to underutilization of video. To address this, we propose a Bootstrapping Video-Language Alignment framework (BoViLA), a self-training method that augments question samples during training process through LLM-based self-questioning and answering, which help model exploit video information and the internal knowledge of LLMs more thoroughly to improve modality alignment. However, low-quality self-generated questions may instead contaminate the performance, especially in the early stages of training, as we have observed in our experiments. To filter bad self-generated questions, we introduce Evidential Deep Learning (EDL) to estimate uncertainty and assess the quality of self-generated questions by evaluating the modality alignment within the context. To the best of our knowledge, this work is the first to explore LLM-based self-training frameworks for modality alignment. We evaluate BoViLA on five strong VideoQA benchmarks, where it outperforms several state-of-the-art methods and demonstrate its effectiveness and generality. Additionally, we provide extensive analyses of the self-training framework and the EDL-based uncertainty filtering mechanism. The code will be made available.

cs.CV↗

arXiv2025

LLM-driven Effective Knowledge Tracing by Integrating Dual-channel Difficulty

Knowledge Tracing (KT) is a fundamental technology in intelligent tutoring systems used to simulate changes in students' knowledge state during learning, track personalized knowledge mastery, and predict performance. However, current KT models face three major challenges: (1) When encountering new questions, models face cold-start problems due to sparse interaction records, making precise modeling difficult; (2) Traditional models only use historical interaction records for student personalization modeling, unable to accurately track individual mastery levels, resulting in unclear personalized modeling; (3) The decision-making process is opaque to educators, making it challenging for them to understand model judgments. To address these challenges, we propose a novel Dual-channel Difficulty-aware Knowledge Tracing (DDKT) framework that utilizes Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) for subjective difficulty assessment, while integrating difficulty bias-aware algorithms and student mastery algorithms for precise difficulty measurement. Our framework introduces three key innovations: (1) Difficulty Balance Perception Sequence (DBPS) - students' subjective perceptions combined with objective difficulty, measuring gaps between LLM-assessed difficulty, mathematical-statistical difficulty, and students' subjective perceived difficulty through attention mechanisms; (2) Difficulty Mastery Ratio (DMR) - precise modeling of student mastery levels through different difficulty zones; (3) Knowledge State Update Mechanism - implementing personalized knowledge acquisition through gated networks and updating student knowledge state. Experimental results on two real datasets show our method consistently outperforms nine baseline models, improving AUC metrics by 2% to 10% while effectively addressing cold-start problems and enhancing model interpretability.

cs.AI↗

arXiv2025

Seed1.5-Thinking: Advancing Superb Reasoning Models with Reinforcement Learning

We introduce Seed1.5-Thinking, capable of reasoning through thinking before responding, resulting in improved performance on a wide range of benchmarks. Seed1.5-Thinking achieves 86.7 on AIME 2024, 55.0 on Codeforces and 77.3 on GPQA, demonstrating excellent reasoning abilities in STEM and coding. Beyond reasoning tasks, the method demonstrates notable generalization across diverse domains. For instance, it surpasses DeepSeek R1 by 8% in win rate on non-reasoning tasks, indicating its broader applicability. Compared to other state-of-the-art reasoning models, Seed1.5-Thinking is a Mixture-of-Experts (MoE) model with a relatively small size, featuring 20B activated and 200B total parameters. As part of our effort to assess generalized reasoning, we develop two internal benchmarks, BeyondAIME and Codeforces, both of which will be publicly released to support future research. Model trial link: https://www.volcengine.com/experience/ark.

cs.CL↗

arXiv2025

Measurement of the dynamic charge susceptibility near the charge density wave transition in ErTe$_3$

A charge density wave (CDW) is a phase of matter characterized by a periodic modulation of the valence electron density accompanied by a distortion of the lattice structure. The microscopic details of CDW formation are closely tied to the dynamic charge susceptibility, $χ(q,ω)$, which describes the behavior of electronic collective modes. Despite decades of extensive study, the behavior of $χ(q,ω)$ in the vicinity of a CDW transition has never been measured with high energy resolution ($\sim$meV). Here, we investigate the canonical CDW transition in ErTe$_3$ using momentum-resolved electron energy loss spectroscopy (M-EELS), a technique uniquely sensitive to valence band charge excitations. Unlike phonons in these materials, which undergo conventional softening due to the Kohn anomaly at the CDW wavevector, the electronic excitations display purely relaxational dynamics that are well described by a diffusive model. The diffusivity peaks around 250 K, just below the critical temperature. Additionally, we report, for the first time, a divergence in the real part of $χ(q,ω)$ in the static limit ($ω\rightarrow 0$), a phenomenon predicted to characterize CDWs since the 1970s. These results highlight the importance of energy- and momentum-resolved measurements of electronic susceptibility and demonstrate the power of M-EELS as a versatile probe of charge dynamics in materials.

cond-mat.str-el↗

arXiv2025

SymTFT Approach to 2D Orbifold Groupoids: `t Hooft Anomalies, Gauging, and Partition Functions

We use the 3D SymTFT approach to study the generalized symmetries and partition functions of 2D CFTs in various orbifolded and fermionic phases. These phases can be realized by the sandwich construction in the associated 3D SymTFTs with different gaped boundaries that encode the data of symmetries in the 2D CFTs. We demonstrate that the gaped boundaries can all be identified with the (fermionic) Lagrangian algebra in the 3D SymTFT, and thus use them to establish webs of dualities of the boundary CFTs in different phases on the level of partition functions. In addition, we introduce the concept of ``para-fermionic Lagrangian algebra" which enables us to construct the partition functions of para-fermionized CFTs on the 2D boundary. Finally, we provide many important examples, including a 3D SymTFT viewpoint on gauging non-invertible symmetries in 2D CFTs.

hep-th↗