论文检索

General Machine Learning · Transfer, Multitask and Meta-learning

Siyang Guo, Junbo Wang, Zibin Zheng

As pre-trained models evolve rapidly, transferring fine-tuning knowledge to updated models without retraining has become a critical challenge. Most existing methods reuse parameter updates, yet the same dataset can induce substantially different updates across base models due to mismatched local loss landscapes, making such transfer unstable. We instead adopt a Bayesian-updating perspective: a base model defines a prior, while fine-tuning contributes a task-update factor that is prior-agnostic, thereby making it feasible to reuse the update across base models. Specifically, we formalize a reusable task-update factor by requiring *invariance across base models* and *a fixed-dimensional parameterization*. Our main theoretical result shows that such reusable factors exist when the variational family is a half-space, and it is already maximal among convex families. In particular, an ideal regime arises when the priors and their Bayesian posteriors remain within a shared exponential family, as it always admits a reusable update factor. Building on this existence, we propose ***B**ayesian Task Update **Transfer*** (*BTransfer*), which extracts a reusable task-update factor from a single fine-tuning run and applies it to a new prior. For deep networks, we implement *BTransfer* with a ``lift–transfer–return'' pipeline: 1) lift model parameters to distributions; 2) transfer the extracted task-update factor in the exponential family distributions; and 3) return the updated posterior distribution to parameter space. Extensive experiments demonstrate that our approach effectively reuses fine-tuning knowledge across models without post-training.

Social Aspects · Accountability, Transparency, and Interpretability

Xinyan Jiang, Ninghao Liu, Di Wang, Lijie Hu

Evaluating LLM reliability via scalar probabilities often fails to capture the structural dynamics of reasoning. We introduce TRACED, a framework that assesses reasoning quality through theoretically grounded geometric kinematics. By decomposing reasoning traces into Progress (displacement) and Stability (curvature), we reveal a distinct topological divergence: correct reasoning manifests as high-progress, stable trajectories, whereas hallucinations are characterized by low-progress, unstable patterns (stalled displacement with high curvature fluctuations). Leveraging these signatures, our probabilistic framework achieves competitive performance and superior robustness across diverse benchmarks. Crucially, TRACED bridges geometry and cognition by mapping high curvature to "Hesitation Loops'' and displacement to ''Certainty Accumulation'', offering a physical lens to decode the internal dynamics of machine thought.

Theory · Reinforcement Learning and Planning

Chenjie Mao, Yi Fan, Ning Zhang, Chongjie Zhang

This paper studies \emph{preference-based reinforcement learning} (PbRL), where agents learn from comparative, trajectory-level feedback rather than numeric rewards. While PbRL has seen rapid empirical and theoretical progress, existing analyses are largely confined to restricted settings and fail to jointly capture the outcome-based and comparison-based nature of preference feedback. We prove that under a broad \emph{general function approximation} framework, PbRL admits a $\sqrt{T}$ regret guarantee. In particular, we introduce a simple and provably efficient algorithm, \emph{Recursive Trajectory-based Preference Q-Learning} (RTPQ), and establish its regret bound while explicitly accounting for the trajectory-level and comparative structure of preferences. Our analysis is characterized by a new complexity measure, the \emph{Dual Episodic Eluder Dimension} (DEED), which quantifies the intrinsic difficulty of PbRL. We show that for linear MDPs, the DEED scales as $\mathcal{O}(dH)$, yielding a regret bound of $\tilde{\mathcal{O}}(dH\sqrt{T}\max(H^{3/2},\,1/\kappa))$, where $\kappa$ is a problem-dependent constant. This bound is near-optimal up to horizon- and problem-dependent factors when compared to standard reward-based linear MDPs. In addition, our framework recovers the best-known regret bounds in the special cases of dueling bandits and standard outcome-based reinforcement learning. Overall, our results provide a general regret guarantee for PbRL with outcome-based preference feedback and broad function approximation.

Kecheng Chen, Ziru Liu, Xijia Tao, Hui Liu, Xinyu Fu, Suiyun Zhang, Dandan Tu, Lingpeng Kong, Rui Liu, Haoliang Li

Diffusion Language Models (DLMs) have recently achieved significant success due to their any-order generation capabilities. However, existing inference methods typically rely on local, immediate-step metrics—such as confidence or entropy—which inherently lack a more reliable perspective, leading to sub-optimal generation quality. To address this, we propose **C**oherent **C**ontextual **D**ecoding (**CCD**), a novel inference framework built upon two core innovations. First, CCD bypasses the potential bias of the single context to leverage historical contexts for approximating the marginal distribution of token prediction, leading to better sequence coherence and the early rejection of sub-optimal paths. More importantly, we demonstrate that this mechanism is theoretically equivalent to modeling the consistency of historical steps via the conditional mutual information between contexts and token predictions. Finally, CCD achieves significantly milder performance degradation under highly parallel decoding scenarios compared to baselines. Empirically, our method achieves a simultaneous enhancement in both inference speed and performance across diverse benchmarks on Dream and LLaDA.

Social Aspects · Fairness

Robin Staab, Jasper Dekoninck, Maximilian Baader, Martin Vechev

Large language models (LLMs) are now widely deployed in user-facing applications, reaching hundreds of millions of users worldwide. Despite their widespread adoption, growing reliance on their outputs raises significant concerns, particularly as users may be exposed to model-inherent biases that disadvantage or stereotype certain groups. However, existing bias benchmarks commonly rely on simple templated prompts or restrictive multiple-choice questions that fail to capture the complexity of real-world user interactions. In this work, we address this gap by introducing a counterfactual framework that automatically generates realistic, open-ended questions for LLM bias evaluation. Through iterative question mutation, our approach further systematically explores areas where models are most susceptible to exhibit biased behavior. Beyond just detecting harmful biases, we also capture increasingly relevant response dimensions, such as asymmetric refusals and explicit bias acknowledgment. Building on this, we construct CAB, a diverse and human-verified benchmark for realistic and nuanced bias evaluations on current frontier LLMs. Our evaluation using CAB highlights the continued need for fairness research by demonstrating that all examined models exhibit persistent biases across certain scenarios.

Social Aspects · Accountability, Transparency, and Interpretability

Guanxu Chen, Dongrui Liu, Tao Luo, Lijie Hu, Qihao Lin, Jing Shao

Large language models (LLMs) are becoming increasingly capable, but the mechanisms of their thinking and decision-making processes remain unclear. Chain-of-thoughts (CoTs) have been commonly utilized to externalize LLMs' thinking, but this strategy fails to accurately reflect LLMs' thinking process. Techniques based on LLMs' hidden representations provide an inner perspective to improve the monitorability of their latent thinking. However, previous methods only try to develop external modules instead of making LLMs themselves easier to monitor. In this paper, we propose a novel method, TELLME, improving the transparency of LLMs and helping monitors identify unsuitable and sensitive behaviors. Furthermore, we showcase the effectiveness of TELLME on detoxification tasks, where LLMs achieve consistent improvement among multimodal test sets, distinct architectures, and varying parameter scales. We further analyze TELLME's improvement on LLMs' generalization ability from both optimal transport theory and empirical perspectives.

General Machine Learning · Transfer, Multitask and Meta-learning

Qun Yang, Enneng Yang, Li Shen, Wei Chen, Long Lan

Continual learning with large pre-trained models offers significant potential for cross-task knowledge accumulation, but faces critical challenges such as catastrophic forgetting and parameter interference, especially when historical data is unavailable. Existing approaches typically rely on sequential fine-tuning or model merging strategies, yet often overlook the impact of loss landscape sharpness and dominant singular value directions, which leads to subspace misalignment and severe knowledge forgetting. In this paper, we propose the Sharpness-Aware Isotropic Merging (SAIM) framework, which introduces targeted optimizations in both the fine-tuning and merging stages to address these issues. Specifically, SAIM consists of two synergistic modules: (1) a Sharpness-Aware Block Coordinate Descent (SA-BCD) optimizer that guides the model toward flatter minima and selectively updates the most task-sensitive parameters, thereby mitigating parameter interference and enhancing robustness; (2) an adaptive isotropic merging algorithm that dynamically balances the singular value spectrum across tasks, effectively preventing the model from overemphasizing any single task direction, maintaining balanced knowledge representation, and improving subspace alignment. Extensive experiments on vision and language benchmarks demonstrate that SAIM achieves 5-10% higher accuracy than existing methods and maintains robust performance as the number of tasks increases. Ablation studies further validate the effectiveness of the SA-BCD fine-tuning strategy in promoting flat minima and reducing parameter interference, as well as its compatibility with various merging approaches.

General Machine Learning · Transfer, Multitask and Meta-learning

Feiyang Deng, Lingfeng Luo, Jiayu Zhou, Kevin He

Accurate survival modeling in biomedical studies is often hindered by rare events, limited effective sample sizes, and settings with limited or partially observed information (e.g., covariates of interest that are difficult or expensive to collect, highly-structured sampling designs, or nuisance parameters omitted by conditioning). Knowledge distillation can leverage external predictive information without sharing individual-level data, but existing approaches are largely built for fully specified likelihoods or probability-based survival models and do not extend to settings where outcome distributions are only partially specified. To address this challenge, we propose a knowledge distillation framework based on a composite-likelihood Kullback–Leibler divergence that aligns teacher and student models within components. Our key insight is that, although composite likelihoods do not define a global outcome distribution, each likelihood component induces a well-defined probability model on its restricted outcome space, enabling a principled KL divergence. Simulation studies and biomedical case studies show improved discrimination and estimation efficiency in rare-event, heterogeneous settings without requiring access to external individual-level data.

General Machine Learning · Causality

Connor T Jerzak, Priyanshi Chandra, Rishi Hazra

We study offline learning of factored stochastic policies over extremely large, combinatorial action spaces and show how standard conjoint data can be used to estimate such policies with asymptotically valid uncertainty under conditions. Conjoint analyses typically report AMCEs by averaging over opponent attributes and thus ignore strategic interdependence. We instead learn \emph{stochastic interventions}---product-of-Categorical policies over factor levels—that (i) optimize expected outcomes in an average-case setting and (ii) extend to a two-player \emph{minimax} (adversarial) setting that realistically captures simultaneous strategic candidate selection. Methodologically, we derive a closed-form solution for the average-case optimizer under two-way interactions with $L_2$ variance regularization, and provide a general gradient-based procedure for richer model classes. Uncertainty from the outcome model propagates asymptotically to both the optimal policy and its value via a Delta method approximation. We further model institutional details (e.g., primaries) inside the minimax objective and introduce a data-driven measure of strategic divergence between parties. On synthetic data, we empirically characterize finite-sample error and coverage as dimensionality and $n$ vary. On a U.S. presidential conjoint, adversarially learned policies produce restricted-equilibrium vote shares that align with historical election ranges in our data, in stark contrast to non-adversarial (averaging) optimizers.

General Machine Learning · Transfer, Multitask and Meta-learning

Mingyue Zeng, De Cheng, Zhipeng Xu, Huaijie Wang, Nannan Wang, Xinbo Gao

Incremental object detection (IOD) aims to extend detectors to new categories while retaining previously acquired knowledge. Existing methods often adopt a class incremental learning perspective, separating feature spaces to sharpen decision boundaries. However, this paradigm conflicts with the inherent object symbiosis in detection, where objects naturally co-occur or occlude one another, creating spatial and semantic dependencies that necessitate a shared feature space. Ignoring these dependencies distorts the shared representations, exacerbates confusion between old and new classes, and accelerates catastrophic forgetting. To address this, we propose Symbiosis-Inspired Knowledge Distillation (SIKD), which explicitly leverages object symbiosis at two complementary levels. Spatial Symbiosis Distillation (SpSD) focuses on symbiotic regions where the old model responds with high overlap to objects in the new task. It preserves generalizable old class cues, suppresses class-specific bias and redundancy, and distills the refined evidence to the new model at matched spatial locations with slot-aligned supervision. Semantic Symbiosis Distillation (SeSD) maintains class level structure by forming confidence weighted prototypes for old classes and aligning their inter class soft ranks over the old class logits, which stabilizes the semantic topology during adaptation. Extensive experiments demonstrate the effectiveness and superiority of the proposed method. Code is available in the supplementary material.

General Machine Learning · Unsupervised and Semi-supervised Learning

Zhi Zhou, Ming Yang, Shi-Yu Tian, Kun-Yang Yu, Lan-Zhe Guo, Yu-Feng Li

Test-time adaptation (TTA) aims to adapt models to maintain reliable performance on non-stationary test streams without requiring labeled data. Despite its empirical success, the learnability of TTA under distributional non-stationarity remains unexplored. A key challenge is lacking of a principled theoretical framework that simultaneously aligns with the TTA objective and captures both continuously evolving distribution shifts and intrinsic information constraints. To address this gap, we propose the first theoretical framework for characterizing the learnability of TTA, introducing the notions of $(\epsilon,\delta)$-Recovery Complexity and $(\epsilon,\rho)$-TTA Learnability. Recovery complexity quantifies the minimal time required for a TTA algorithm to recover to a target excess risk following a distribution shift, and is further generalized to define $(\epsilon,\rho)$-TTA Learnability, which measures the long-term reliability of TTA algorithms over non-stationary data streams. Within this framework, we introduce a novel temporally dependent discrete surrogate method that models complex non-stationary test streams, enabling a unified and tractable analysis of both gradual and abrupt distribution shifts. We derive order-wise matching lower and upper bounds on recovery complexity through information-theoretic and optimization-based analysis. Our results uncover fundamental limits of TTA, highlight the intrinsic adaptivity-information trade-off of TTA algorithms, and provide the first unified learnability guarantees that go beyond regret-based perspectives.

Reinforcement Learning · Online

Yiting He, Zhishuai Liu, Pan Xu

Empirical successes show that regularization improves the stability and efficiency of reinforcement learning (RL), with applications in robotics and post-training of large language models. Yet, theoretical analyses of regularized Markov decision processes (MDPs) have mostly been confined to the standard RL setting. In this work, we investigate regularized MDPs through the lens of robust RL. We introduce a doubly regularized MDP framework that combines policy and dynamics regularization to enable robust policy learning against reward and dynamics perturbations. Within this framework, we develop an optimism-based online algorithm and provide the first finite-sample regret guarantees in both tabular and rich-observation settings, where the state-action space may be continuous. Our results show that algorithms for doubly regularized MDPs are as sample-efficient as well-studied robust MDP algorithms, while additionally benefiting from the flexibility of soft policies. Finally, we use experiments to demonstrate that our approach efficiently and effectively handles function approximation and exploration in large state-action spaces, achieving robust performances.

Applications · Language, Speech and Dialog

Ke Xue, Rongfei Fan, Kai Li, Shanping Yu, Puning Zhao, Jianping An

Diffusion models have recently set new benchmarks in Speech Enhancement (SE). However, most existing score-based models treat speech spectrograms merely as generic 2D images, applying uniform processing that ignores the intrinsic structural sparsity of audio, which results in inefficient spectral representation and prohibitive computational complexity. To bridge this gap, we propose **DVPD**, an extremely lightweight **D**ual-**V**iew **P**redictive **D**iffusion model, which uniquely exploits the dual nature of spectrograms as both visual textures and physical frequency-domain representations across both training and inference stages. Specifically, during training, we optimize spectral utilization via the Frequency-Adaptive Non-uniform Compression (FANC) encoder, which preserves critical low-frequency harmonics while pruning high-frequency redundancies. Simultaneously, we introduce a Lightweight Image-based Spectro-Awareness (LISA) module to capture features from a visual perspective with minimal overhead. During inference, we propose a Training-free Lossless Boost (TLB) strategy that leverages the same dual-view priors to refine generation quality without any additional fine-tuning. Extensive experiments across various benchmarks demonstrate that DVPD achieves state-of-the-art performance while requiring only **35** of the parameters and **40%** of the inference MACs compared to SOTA lightweight model, PGUSE. These results highlight DVPD's superior ability to balance high-fidelity speech quality with extreme architectural efficiency. Code and audio samples are available at the anonymous website: https://anonymous.4open.science/r/dvpd_demo-E630

Deep Learning · Large Language Models

Dexuan Xu, Jieyi Wang, Shijie Li, Hanpin Wang, Yongzhi Cao, Yu Huang

Lifelong multimodal knowledge editing allows vision language models to continuously adapt to dynamic updates to avoid catastrophic forgetting. To mitigate interference between sequential updates, recent paradigms have shifted towards modular parameter isolation. However, this strategy faces a critical scalability bottleneck: accumulating dense parameter blocks can lead to excessive memory growth, and managing these independent modules often uses decoupled routing mechanisms, resulting in architectural redundancy. To address this issue, we propose **TIME** (**T**ensor-Factorized **I**ntrinsic **M**ixture-of-**E**xperts), a unified framework harmonizing parameter efficiency with structural self-routing. TIME parameterizes each knowledge edit as a compact CP-decomposed tensor, significantly reducing complexity compared to low-rank matrices. Furthermore, departing from auxiliary semantic retrievers, we introduce an intrinsic routing mechanism that utilizes the tensor's input factors to directly define the active subspace, effectively enabling expert parameters to serve simultaneously as the routing logic. Extensive experiments demonstrate that TIME achieves state-of-the-art performance on lifelong editing benchmarks while successfully reducing memory usage and inference latency.

Deep Learning · Large Language Models

Ziheng Ren, Zhanming Shen, Hao Wang, Ning Liu, You Song

Federated fine-tuning of Large Language Models faces severe statistical heterogeneity. However, existing model-level defenses often overlook the root cause: intrinsic data distribution mismatches. In this work, we first establish Federated Self-Distillation (FedSD) as a fundamental and potent strategy. By projecting client representations into a smoothed ``model-understanding space,'' FedSD alone serves as a universal booster, demonstrating superior performance over conventional algorithms. Despite its success, we identify a subtle trade-off termed the Rewrite Paradox---unconstrained self-distillation can inadvertently increase hallucinations and redundancy. To refine this paradigm, we further propose FedSDR (Federated Self-Distillation with Rectification), the ultimate reinforced framework. It augments FedSD with a dual-stream mechanism: a local LoRA-S (Smoothing) branch to implicitly absorb heterogeneity via distilled data, and a parallel global LoRA-R (Rectification) branch anchored to raw data to enforce factual correctness. By selectively aggregating only LoRA-R, FedSDR yields a globally aligned and faithful model. Extensive experiments verify its superior performance.

Social Aspects · Safety

David Guzman Piedrahita, Dave Banerjee, Changling Li, Terry Zhang, Kevin Blin, Samuel Simko, Punya Pandey, Irene Strauss, Rada Mihalcea, Bernhard Schölkopf 等

Sociopolitical AI risks are threats to collective self-determination: a society's capacity to articulate its interests and realize them through institutions. We argue that sociopolitical AI risks emerge when general-purpose AI systems are integrated into society in ways that disproportionately amplify the scale, speed, and opacity of institutional operations, thereby degrading their capacity to function. Unlike model-level harms (toxicity, bias, discrimination), sociopolitical risks arise from widespread deployment rather than individual outputs. And unlike existential risks involving loss of control or complete labor automation, they manifest with current AI capabilities where AI augments rather than replaces human activity. In this position paper, we analyze how AI alters the conditions of governance: flooding government agencies with paralyzing volumes of input, concentrating control of infrastructure that threatens sovereignty, and flattening public debate into artificial agreement while reinforcing existing biases.

Applications · Health / Medicine

Changshuo Liu, Wu Junran, Zhongle Xie, Wenqiao Zhang, Kaiping Zheng, Jiaqi Zhu, Qingpeng Cai, Gene Anne Ooi, Marcus CJ Tan, Jianwei Yin 等

Generative AI is reshaping healthcare, yet most existing advances rely on hospital-grade devices, which limits their accessibility and potential for health management outside clinical settings. With the proliferation of portable devices and telemedicine, healthcare is shifting toward home-based Diagnosis-It-Yourself (DIY) care. Despite this promise, several distinctive challenges remain: (i) home-collected data are heterogeneous, exacerbated by the absence of standardized large-scale datasets; (ii) models require adaptation to variable task demands and evolving individual conditions; (iii) the broad spectrum of home care tasks lacks a unified benchmark for systematic evaluation. In this paper, we present **DIYHealth Suite**, a comprehensive framework designed to address these challenges through a tailored dataset, model, and benchmark. We first curate **DIYHealth-900K**, a large-scale multimodal dataset capturing diverse real-world home care scenarios. Building on this, we propose **DIYHealthGPT**, an adaptive foundation model for home-based health management, powered by the novel Hybrid Hyper Low-Rank Adaptation technique. Finally, we establish **DIYHealthBench**, the first benchmark to evaluate foundation models on home care tasks. Extensive experiments demonstrate that DIYHealthGPT delivers state-of-the-art performance over both general-purpose and medical-specific baselines on 11 home care tasks in both open-QA and closed-QA settings, laying the groundwork for the next generation of personalized health management at home.

Reinforcement Learning · Everything Else

Jintao Li, Maowen Tang, Yongji Long, Weixuan Liu, Yanlang Zheng, Sicheng He, Ao-Jin Li, Shui Yu, Yun Li

Constrained multi-objective reinforcement learning aims to discover a diverse set of feasible trade-offs, yet scalarization and signed, normalized group-relative advantages can be brittle under objective-scale drift, near-ties, and feasibility scarcity. We propose constrained projection policy optimization (CoPro), which alternates between an E-step moment projection and an M-step policy projection. In the E-step, we solve a Kullback-Leibler (KL)-regularized, moment-constrained projection over each sampled group to compute a nonnegative reweighting distribution (q*) that promotes feasible Pareto-front (PF) progress, preserves feasibility anchors, and suppresses ambiguous near-ties. This E-step admits a closed-form exponential-family solution and guarantees strictly positive probability mass on feasible anchors whenever feasible candidates appear in the group. In the M-step, we project the policy toward q* via weighted maximum likelihood with a trust-region regularizer, yielding a PF-aligned update direction from comparisons without hand-crafted reward shaping. Empirically, CoPro improves feasible PF quality and robustness on constrained multi-objective benchmarks for large language model tool use and analog circuit design tasks. Code is available at https://anonymous.4open.science/r/CoPro-8A95/README.md.

Social Aspects · Privacy

Anneliese Riess, Felipe Gomez, Flavio Calmon, Julia Schnabel, Georgios Kaissis

We prove the conjecture stated in Appendix F.3 of Zhu et al.: among all conversion rules that map a Rényi Differential Privacy (RDP) profile $\tau \mapsto \rho(\tau)$ to a valid hypothesis-testing trade-off $f$ (or equivalently, an $(\varepsilon,\delta)$-Differential Privacy curve), the rule based on the intersection of single-order RDP privacy regions is optimal. This optimality holds simultaneously for all valid RDP profiles and for all Type I error levels $\alpha$. Concretely, we show that in the space of trade-off functions, the tightest possible bound is $f_{\rho(\cdot)}(\alpha) = \sup_{\tau \geq 0.5} f_{\tau,\rho(\tau)}(\alpha)$: the pointwise maximum of the single-order bounds for each RDP privacy region. Our proof unifies and sharpens the insights of Balle et al., Asoodeh et al., and Zhu et al.. Our analysis relies on a precise geometric characterization of the RDP privacy region, leveraging its convexity and the fact that its boundary is determined exclusively by Bernoulli mechanisms. Our results establish that the "intersection-of-RDP-privacy-regions" rule is not only valid, but optimal: no other black-box conversion can uniformly dominate it in the Blackwell sense, marking the fundamental limit of what can be inferred about a mechanism's privacy solely from its RDP guarantees.

General Machine Learning · Unsupervised and Semi-supervised Learning

Hasan Sabri Melihcan Erol, Suat Evren, Oktay Ozel, Alexander Morgan, Jongha (Jon) Ryu, Lizhong Zheng

Contrastive learning is often trained with the InfoNCE loss, which uses a softmax over similarities to make the positive pair score higher than many negatives. Beyond its connection to mutual information, this softmax link has a precise probabilistic meaning: it is the maximum likelihood objective of a discrete choice model Plackett Luce with Gumbel noise. We show that this implicit noise model can be systematically wrong in modern settings where similarities are bounded, such as cosine normalized embeddings. In the bounded regime, the most competitive negatives pile up near the score ceiling, and extreme value theory predicts Weibull rather than Gumbel behavior for these extremes. We confirm this prediction empirically by measuring Weibull style tail behavior in the hardest negatives throughout InfoNCE training. Motivated by this mismatch, we propose WEINCE, a simple modification of InfoNCE that targets the extreme score regime directly. Across standard benchmarks and backbone architectures, WEINCE improves downstream linear evaluation over InfoNCE with minimal changes to existing training pipelines, showing that modeling the geometry of extremes can yield stronger contrastive representations.