论文检索

Deep Learning · Generative Models and Autoencoders

Xi Wang, Wenbo Lu, Shenji Wan

Generative Flow Networks (GFlowNets) enable fine-tuning large language models to approximate reward-proportional posteriors, but they remain prone to mode collapse, manifesting as prefix collapse and length bias. We attribute this to two factors: (i) weak credit assignment to early prefixes, and (ii) biased replay that induces a shifted, non-representative training flow distribution. We propose Rooted absorbed prefix Trajectory Balance (RapTB), an objective that anchors subtrajectory supervision at the root and propagates terminal rewards to intermediate prefixes via absorbed suffix-based backups, providing dense prefix-level learning signals. To mitigate replay-induced distribution shift, we further introduce SubM, a submodular replay refresh strategy that promotes both high reward and diversity. Empirically, on tasks such as molecule generation with LLM using SMILES strings, RapTB combined with SubM consistently improves optimization performance and molecular diversity while preserving high validity.

Deep Learning · Generative Models and Autoencoders

AN HUANG, Junggab Son, Zuobin Xiong

Diffusion models have become the foundation of modern generative systems, with most research focusing primarily on improving generation efficiency and output quality. The timestep embedding component is a crucial part of the diffusion pipeline, which provides a temporal conditioning signal to the denoising network, enabling it to adapt its predictions across different noise levels throughout the process. Despite their potential to contain substantial information, timestep embeddings remain underexplored in current research, especially for security risks and reliable provenance. To fill this gap, we introduce **Shadow Timestep Embedding (STE)**, a novel mechanism that investigates the underutilized temporal space for malicious information injection into diffusion models. In particular, when zooming in on the timestep embedding space, we find that different timesteps exhibit distinct representational capabilities that can encode side-channel information. Moreover, such encoded information can be utilized for attack and defense purposes through the scheduler interface. We present a theoretical analysis of timestep embeddings as position-encoding mappings and derive a mutual coherence evaluation that explains the separability of disjoint timestep intervals. Our findings reveal the diffusion model's timestep as a powerful side channel for carrying dedicated information, motivating new directions for adversarial generative modeling by understanding the temporal dimension.

Deep Learning · Large Language Models

Yujie Zheng, Zhuo Li, Shengtao Zhang, Jiaqian Wang, Junjie Sheng, Junchi Yan, Weinan Zhang, Ying Wen, Bo Tang, Muning Wen

Deploying Large Language Models to data-scarce programming domains poses significant challenges, particularly for kernel synthesis on emerging Domain-Specific Architectures where a "Data Wall" limits available training data. While models excel on data-rich platforms like CUDA, they suffer catastrophic performance drops on data-scarce ecosystems such as NPU programming. To overcome this cold-start barrier without expensive fine-tuning, we introduce Evokernel, a self-evolving agentic framework that automates the lifecycle of kernel synthesis from initial drafting to continual refining. Our method addresses this by formulating the synthesis process as a memory-based reinforcement learning task. Through a novel value-driven retrieval mechanism, it learns stage-specific Q-values that prioritize experiences based on their contribution to the current objective—whether bootstrapping a feasible draft or iteratively refining latency. Furthermore, by enabling cross-task memory sharing, the agent generalizes insights from simple to complex operators. By building an NPU variant of KernelBench and evaluating on it, \ourmethod improves frontier models' correctness from 11.0% to 83.0% and achieves a median speedup of 3.60x over initial drafts through iterative refinement. This demonstrates that value-guided experience accumulation allows general-purpose models to master the kernel synthesis task on niche hardware ecosystems.

Social Aspects · Privacy

Sander De Coninck, Sam Leroux, Pieter Simoens

The adoption of computer vision to drive industrial efficiency and safety creates a persistent tension between operational utility and worker surveillance. Current privacy measures, such as post-hoc blurring, are fundamentally flawed: they depend on the error-prone detection of sensitive attributes and treat privacy as a subtractive process. We posit that industrial computer vision must shift from "hiding secrets'' to verifiable data minimization. We advocate for a design paradigm of architecturally constrained inference, formalized through information-theoretic principles, where the sensing pipeline is optimized to capture only the features necessary for a specific task (e.g., pose estimation). This provably constrains the information available for unauthorized inferences (e.g., identification), decoupling privacy from detection accuracy and reducing reliance on sensitive attribute supervision. We outline an implementation path using modular edge processing and trusted execution environments to enable verifiable, hardware-rooted attestations of task-bound processing, and argue that verifiable purpose limitation should be a prerequisite for responsible industrial AI.

Deep Learning · Generative Models and Autoencoders

Mingyu Wang, Wei Jiang

Recent advances in generative distillation have shown strong potential in constructing high quality surrogate datasets within a fraction of the time required by optimization-based approaches. However, most existing generative solutions rely on diffusion models, which suffer from two limitations. (i) Indirect matching objectives. Their sequential denoising process makes it difficult to directly match representative prototypes. (ii) Target-agnostic generation. The generation process is often decoupled from the target task, causing the synthesized samples to drift from the desired distribution. Building on this insight, We propose ProtoVAR, a prototype-guided visual autoregressive framework. Instead of relying on latent space, ProtoVAR uses the coarse-to-fine next-scale prediction of Visual AutoRegressive (VAR) modeling to maintain semantic consistency during generation. By injecting multi-scale class prototypes, ProtoVAR enforces clear representativeness constraints while preserving diversity. A pool-based selector further distills the prototype-guided outputs into a compact, task-aligned surrogate dataset. Extensive experiments show that ProtoVAR achieves state-of-the-art performance with comparable or lower computational cost than diffusion-based distillation.

Optimization · Zero-order and Black-box Optimization

Yiyi Zhu, Yaolin Wen, Xiang Xia, Xin An, Hanyi Si, Xiang Shu, Yangde Fu, Liang Dou, Hong Qian

Multi-objective optimization (MOO) has emerged as a powerful approach to solving complex optimization problems involving multiple objectives. In many practical scenarios, function evaluations are unavailable or prohibitively expensive, necessitating optimization solely based on a fixed offline dataset. In this setting, known as offline MOO, the goal is to find out the Pareto set without access to the true objective functions. This setting suffers from the out-of-distribution (OOD) issue, where the surrogate model is not accurate for unseen designs. Due to the OOD issue, surrogate errors may cause the optimizer to select solutions that do not lie on the true Pareto front and are biased toward its extremes. To address this, this paper proposes Diversity-driven Offline Multi-Objective Optimization (DOMOO), which aims to find out a diverse and high-quality set of solutions. Firstly, DOMOO incorporates an accumulative risk control module that estimates the potential risk of candidate solutions and alleviates the OOD issue between the training data and the generated solutions. In addition, a nested Pareto set learning (PSL) strategy is proposed to jointly learn preference and PSL parameters, then optimize them, enabling adaptation to diverse Pareto front geometries. To further enhance solution quality, we design a diversity-driven selection strategy that extracts a representative and well-distributed set of final solutions. To achieve this diversity-driven selection strategy, we propose $\text{IGD}\_{\text{offline}}$, a tailored indicator for the offline setting that considers both diversity and convergence, and avoids the bias of hypervolume indicator. Extensive experiments on synthetic and real-world benchmarks, such as neural architecture search, show that, on average across benchmarks, DOMOO achieves a 1.44× improvement in convergence and diversity over comparable methods.

Social Aspects · Security

Anran Zhu, Zhengli Shi, Chende Zheng, Chenhao Lin, Zhengyu Zhao, Le Yang, Chong Zhang, Shuai Liu, Chao Shen

With the rapid advancement of high-fidelity video generation models, robust AI-generated video (AIGV) detection has become increasingly needed. While most AIGV detection methods operate in the decoded pixel domain, we observe that detection in the pixel domain inevitably entangles task-irrelevant semantic information, leading to substantial semantic redundancy and extensive redundant computation, while overlooking free-to-use signals in compressed bitstreams. In particular, motion vectors and residuals directly encode temporal and spatial generative artifacts but remain largely underexplored. To address these issues, we propose a unified framework for **S**patio-**T**emporal **RE**sidual and **A**rtifact **M**ining, namely **STREAM**, which enables AIGV detection directly from compressed bitstreams. **STREAM** leverages I-frames, motion vectors, and residual errors to capture spatiotemporal artifacts that are typically smoothed out by decompression filters. In particular, we design a lightweight network with a motion-guided alignment module and a gated fusion mechanism, enabling adaptive fusion of spatial artifacts and nonlinear temporal dynamics. Extensive experimental results demonstrate that **STREAM** achieves SOTA performance with an mAP of 0.965, with 2.5× faster inference than previous SOTA baselines.

Applications · Computer Vision

Zihui Zhang, Zhixuan Sun, Yafei YANG, Jinxi Li, Jiahao Chen, Bo Yang

We address the challenging task of 3D object segmentation in complex scene point clouds without relying on any scene-level human annotations during training. Existing methods are typically constrained to identifying simple objects, primarily due to insufficient object priors in the learning process. In this paper, we present FoundObj, a novel framework featuring a superpoint-based object discovery agent that incrementally merges suitable neighboring superpoints, guided by our innovative semantic and geometric reward modules. These modules synergistically leverage semantic and geometric priors from self-supervised 2D/3D foundation models, providing complementary feedback to the object discovery agent and enabling robust identification of multi-class objects through reinforcement learning. Extensive experiments on diverse benchmarks demonstrate that our approach consistently outperforms existing baselines. Notably, our method exhibits strong generalization in zero-shot and long-tail scenarios, underscoring its potential for scalable, label-free 3D object segmentation.

Optimization · Stochastic

Zitao Song, Cedar Site Bai, Zhe Zhang, Brian Bullins, David Gleich

Adaptive methods like Adam have become the *de facto* standard for large-scale vector and Euclidean optimization due to their coordinate-wise adaptation with a second-order nature. More recently, matrix-based spectral optimizers like Muon (Jordan et al., 2024b) show the power of treating weight matrices as matrices rather than long vectors. Linking these is hard because many natural generalizations are not feasible to implement, and we also cannot simply move the Adam adaptation to the matrix spectrum. To address this, we reformulate the AdaGrad update and decompose it into a variance adaptation term and a scale-invariant term. This decoupling produces **DeVA** (**De**coupled **V**ariance **A**daptation), a framework that bridges between vector-based variance adaptation and matrix spectral optimization, enabling a seamless transition from Adam to adaptive spectral descent. Extensive experiments across language modeling and image classification demonstrate that DeVA consistently outperforms state-of-the-art methods such as Muon and SOAP (Vyas et al., 2024), reducing token usage by around 6.6\%. Theoretically, we show that the variance adaptation term effectively improves the blockwise smoothness, facilitating faster convergence.

Applications · Chemistry, Physics, and Earth Sciences

Kiyoung Seong, Sungsoo Ahn, Sehui Han, Changyoung Park

Crystal modeling spans a family of conditional and unconditional generation tasks across different modalities, including crystal structure prediction (CSP) and *de novo* generation (DNG). While recent deep generative models have shown promising performance, they remain largely task-specific, lacking a unified framework that shares crystal representations across different generation tasks. To address this limitation, we propose *Multimodal Crystal Flow (MCFlow)*, a unified multimodal flow model that realizes multiple crystal generation tasks as distinct inference trajectories via independent time variables for atom types and crystal structures. To enable multimodal flow in a standard transformer model, we introduce a composition- and symmetry-aware atom ordering with hierarchical permutation augmentation, injecting strong compositional and crystallographic priors without explicit structural templates. Experiments on the MP-20 and MPTS-52 benchmarks show that MCFlow achieves competitive performance against task-specific baselines across multiple crystal generation tasks. Our code and inference trajectories are available at [https://anonymous.4open.science/r/mcflow-46E4](https://anonymous.4open.science/r/mcflow-46E4).

Social Aspects · Accountability, Transparency, and Interpretability

Simon Schrodi, Julian Schur, Max Argus, Thomas Brox

Concept-based models like Concept Bottleneck Models (CBMs) have garnered significant interest for improving model interpretability by first predicting human-understandable concepts before mapping them to the output classes. Early approaches required costly concept annotations. To alleviate this, recent methods utilized large language models to automatically generate class-specific concept descriptions and learned mappings from a pretrained black-box model’s raw features to these concepts using vision-language models. However, these approaches assume prior knowledge of which concepts the black-box model has learned. In this work, we discover the concepts encoded by the model through unsupervised concept discovery techniques instead. We further leverage a simple input-dependent concept selection mechanism that dynamically retains a sparse set of relevant concepts of each input, enhancing both sparsity and interpretability. Our approach not only improves downstream performance, but also needs significantly fewer concepts for accurate classification. Lastly, we show how large vision-language models can guide the editing of our models' weights to correct model errors.

Deep Learning · Large Language Models

Hyesung Jeon, Hyeongju Ha, jae-joon kim

Role specialization in multi-LLM agent systems is often realized via multi-LoRA, where agents share a pretrained backbone and differ only through lightweight adapters. Despite sharing base model weights, each agent independently builds and stores its own KV cache for the same long, tool-augmented trajectories, incurring substantial memory and compute overhead. Existing KV cache sharing methods largely overlook this multi-LoRA setting. We observe that, across agents, cache differences are dominated by adapter outputs, while activations from the shared pretrained backbone remain highly similar. Based on this observation, we propose LRAgent, a KV cache sharing framework for multi-LoRA agents that decomposes the cache into a shared base component from the pretrained weights and an adapter-dependent component from LoRA weights. LRAgent reduces memory overhead by sharing the base component and storing the adapter component in its inherent low-rank form, and further reduces compute overhead, enabled by shared-$A$ multi-LoRA architectures, by also sharing the low-rank cache and avoiding redundant computations for contexts already processed by other agents. To efficiently reconstruct adapter contributions at runtime, we introduce Flash-LoRA-Attention, a kernel that reorders attention computation to avoid materializing the low-rank cache to full dimension. LRAgent achieves throughput and time-to-first-token latency close to fully shared caching, while preserving accuracy near the non-shared caching baseline across agentic question-answering benchmarks.

General Machine Learning · Supervised Learning

Guangzheng Hu, Patricia Menendez Galvan, Feng Liu, Mingming Gong, Guanghui Wang, Liuhua Peng

Federated learning has emerged as the foremost approach for decentralized model training with privacy preserving. The global class imbalance and cross-client data heterogeneity naturally coexist, and the mismatch between local and global imbalances exacerbates the performance degradation of the aggregated model. The agnosticism of global class distribution poses significant challenges for data-level methods, especially under extreme conditions with severe class absence across clients. In this paper, we propose FedReLa, a novel data-level approach that tackles the coexistence of data heterogeneity and class imbalance in federated learning. By re-labeling samples with a feature-dependent label re-allocator, FedReLa corrects biased global decision boundaries without requiring knowledge of the global class distribution. This modular, model-agnostic approach can be integrated with algorithmic methods to deliver consistent improvements without additional communication overhead. Through extensive experiments, our method significantly improves the accuracy of minority classes and the overall accuracy on stepwise-imbalanced and long-tailed datasets, outperforming the previous state of the art.

Applications · Health / Medicine

Akash Ghosh, Srivarshinee Sridhar, Raghav Ravi, Muhsin Muhsin, Sriparna Saha, Chirag Agarwal

Integrating language models (LMs) in healthcare systems holds great promise for improving medical workflows and decision-making. However, a critical barrier to their real-world adoption is the lack of reliable evaluation of their trustworthiness, especially in multilingual healthcare settings. Existing LMs are predominantly trained in high-resource languages, making them ill-equipped to handle the complexity and diversity of healthcare queries in mid- and low-resource languages, posing significant challenges for deploying them in global healthcare contexts where linguistic diversity is key. In this work, we present \textsc{Clinic}, a \textbf{C}omprehensive Mu\textbf{l}tilingual Benchmark to evaluate the trustworth\textbf{i}ness of la\textbf{n}guage models \textbf{i}n health\textbf{c}are. \name systematically benchmarks LMs across five key dimensions of trustworthiness: truthfulness, fairness, safety, robustness, and privacy, operationalized through 18 diverse tasks, spanning 15 languages (covering all the major continents), and encompassing a wide array of critical healthcare topics like disease conditions, preventive actions, diagnostic tests, treatments, surgeries, and medications. Our extensive evaluation reveals that LMs struggle with factual correctness, demonstrate bias across demographic and linguistic groups, and are susceptible to privacy breaches and adversarial attacks. By highlighting these shortcomings, \name lays the foundation for enhancing the global reach and safety of LMs in healthcare across diverse languages.

Applications · Chemistry, Physics, and Earth Sciences

Chenghao Jia, Mengdi Liu, Hong Chang, Shiguang Shan, Xilin Chen

Elucidating molecular structures from spectra is a foundational problem in chemical and materials characterization, yet remains challenging due to spectral ambiguity and the vast molecular space. Although recent diffusion-based generators show strong promise for spectra-conditioned elucidation, existing methods struggle to learn robust spectra-structure relationships from limited paired data when relying solely on global spectral representation. Moreover, the repeated full sampling inference strategy incurs substantial computation overhead. To address these limitations, we propose MAST, a Motif-Augmented diffusion framework with Search Tree, for joint 2D-3D spectroscopic molecular structure elucidation. MAST introduces explicit, interpretable motif priors as intermediate evidences throughout denoising, reducing conditional ambiguity and facilitating spectra-conditioned optimization. We further cast diffusion sampling as reward-guided tree search to prioritize high-reward denoising trajectories, yielding a compact set of spectra-consistent candidates under limited budgets. On the QM9S multi-spectra benchmark, MAST achieves 94.89% exact recovery and improves 3D fidelity, while preserving high chemical validity and stability.

Deep Learning · Large Language Models

Shashwat Goel, Rishi Hazra, Dulhan Jayalath, Timon Willi, Parag Jain, Shen, Ilias Leontiadis, Francesco Barbieri, Yoram Bachrach, Jonas Geiping 等

AI co-scientists are emerging as a useful tool for human researchers, with a crucial ability being proposing a research plan for a given research goal. In this work, we study how to train language models that generate better research plans by leveraging the vast corpus of existing research papers. To collect diverse training data, we automatically extract research goals and goal-specific grading rubrics from papers across domains. We then train models for research plan generation via reinforcement learning, with a frozen copy of the initial policy acting as the grader, using the rubrics to evaluate plans generated by the training policy. To validate this approach, we conduct a human study for machine learning research goals spanning 225 expert hours. The experts prefer plans generated by our finetuned Qwen3-30B-A3B model over the initial model for 70% goals, and over Grok-4-Thinking for 59.6% goals. To assess generality, we also extend our approach to goals from medical papers, and recent arXiv preprints, evaluating with a jury of frontier models. Our finetuning yields 12-22% relative improvements and significant cross-domain generalization, proving effective even in problem settings like medical research where execution feedback is infeasible. Overall, we demonstrate the potential of a scalable training recipe as a step towards improving general AI co-scientists.

Deep Learning · Large Language Models

Dayuan Zhao, Shengcao Cao, Yu-Xiong Wang, Liang-Yan Gui

Latent reasoning has emerged as a powerful alternative to text-based Chain-of-Thought (CoT), offering significant gains in computational efficiency by compressing verbose reasoning into compact embeddings. However, compressing reasoning into the latent space renders the thinking opaque, hindering its interpretability. Current methods present a stark trade-off: they either function as unexplainable “black boxes” (e.g., Coconut), where the latent reasoning is not human-readable, or rely on separate post-hoc decoders for explainability (e.g., Heima), introducing architectural overhead and decoupling the explanation from the actual reasoning process. In this work, we present a unified framework for Self-Explainable Latent Reasoning (SELR) that trains a single model to perform efficient and inherently explainable latent reasoning. Our core contribution is a novel multi-task training objective that optimizes for two goals simultaneously: (1) an Answer Loss that optimizes the latent reasoning trajectory to produce accurate final answers, and (2) a CoT Loss that explicitly trains the same model to decode its own latent representations back into human-understandable reasoning steps. This design ensures that generated latent representations are both task-effective and semantically interpretable, eliminating the need for external decoders. We validate the effectiveness of SELR on both Large Language Models (LLMs) and Vision-Language Models (VLMs), demonstrating that SELR achieves superior token efficiency and accuracy compared to baselines, while uniquely providing self-contained explainability without auxiliary models.

Deep Learning · Large Language Models

Shuai Shao, Yixiang Liu, Bingwei Lu, Weinan Zhang

In recent years, LLM-based multi-agent systems (MAS) have advanced rapidly, using a router to decompose tasks and delegate subtasks to specialized agents. A natural way to expand capability is to **scale up the agent pool** by continually integrating new functional agents or tool interfaces, but naive expansion can trigger **performance collapse** when the router cold-starts on newly added, heterogeneous, and unreliable agents. We propose **MonoScale**, an expansion-aware update framework that proactively generates a small set of agent-conditioned familiarization tasks, harvests evidence from both successful and failed interactions, and distills it into auditable natural-language memory to guide future routing. We formalize sequential augmentation as a contextual bandit and perform trust-region memory updates, yielding a monotonic non-decreasing performance guarantee across onboarding rounds. Experiments on GAIA and Humanity's Last Exam show stable gains as the agent pool grows, outperforming naive scale-up and strong-router fixed-pool baselines.

Deep Learning · Large Language Models

Yaoyou Fan, Chao Zhang, Xiaoyu Tan, Chenxing Sun, Yu Yuan, Haoyu Feng, Lu Pan, Ke Zeng, Xunliang Cai

Supervised Fine-Tuning (SFT) with Negative Log-Likelihood (NLL) remains the standard post-training paradigm for Large Language Models, yet it imposes an excessive penalty on low-probability target tokens. This focus forces the model to prioritize minimizing the loss of difficult samples over optimizing the overall quality of the generation, often leading to unwarranted overconfidence. On the other hand, alternatives like Dynamic Fine-Tuning (DFT) suffer from vanishing gradients on these tokens, which severely hinders the acquisition of new concepts. To bridge this gap, we propose **SAFT** **S**pectrum-**A**daptive **F**ine-**T**uning), a unified framework that interpolates between the aggressive learning signal of NLL and the robust nature of probability-weighted optimization. By adaptively balancing these objectives, SAFT effectively mitigates outlier sensitivity without sacrificing learning efficiency. Empirically, our method achieves state-of-the-art performance on mathematical reasoning benchmarks, demonstrating superior generalization on out-of-distribution tasks. Our anonymized code is available at https://anonymous.4open.science/r/SAFT-9FEB.

Deep Learning · Theory

Akira Sakai, Yuma Ichikawa

Sub-bit model compression seeks storage below one bit per weight, where the sign bit becomes a fixed-cost bottleneck as magnitudes are aggressively compressed. Across Transformers, CNNs, and MLPs, learned sign matrices resist low-rank compression and are spectrally indistinguishable from i.i.d. Rademacher baselines. Despite this apparent randomness, most weights keep their initialization signs, with flips occurring mainly through rare near-zero boundary crossings, **suggesting that the randomness in sign patterns is largely inherited from initialization.** We formalize this behavior with *sign lock-in theory*, a stopping-time analysis of sign flips under SGD noise. Under bounded updates and a rare re-entry condition for a small neighborhood around zero, the number of effective sign flips exhibits a geometric tail. Building on this mechanism, we introduce a gap-based initialization and a lightweight outward-drift regularizer that reduces the effective flip rate to approximately $10^{-3}$ with only about a one-point increase in perplexity.