论文检索

Reinforcement Learning · Deep RL

Chen Jia

We propose to model text generation as a continuous-time latent dynamical process, where token generation is formulated as a Markov Decision Process whose internal state evolves via a neural ODE. This formulation bridges discrete token sequences and continuous semantic evolution, providing a theoretically grounded approach for coherent long-range generation. The framework is optimized via reinforcement learning, maximizing a composite objective that integrates task-specific rewards with knowledge distillation from a powerful pre-trained language model. Experiments demonstrate that our method, Continuous-Time Latent Language Model (CT-LLM), outperforms discrete baselines in generation coherence and long-context performance, offering a new paradigm for fluid and controllable language generation.

General Machine Learning · Everything Else

Yingpeng Tang, Zhuang Qi, Xiaoli Tang, Wei Zhuo, Sheng-Jun Huang, Han Yu

With the advance of federated learning (FL) in privacy-sensitive domains such as healthcare, finance, and mobile intelligence, the need for efficient and robust training becomes increasingly urgent. Communication bottlenecks, heterogeneous client distributions, and fairness requirements make it essential to select the “right” data and features for model training. Yet existing FL research often addresses feature selection and data selection separately, ignoring their interplay in real-world high-dimensional and noisy datasets, leading to suboptimal performance. In this paper, we propose a unified framework for data and feature selection by formulating the problem as a generalized CUR decomposition problem. We introduce FedGCUR, a practical framework that integrates a federated column-pivoted QR (FedCPQR) decomposition routine with per-silo row selection. Specifically, FedCPQR is designed to securely compute a global pivot order without exposing raw data, while FedGCUR leverages this to jointly select shared features and silo-specific samples. We prove that FedCPQR produces exactly the same decomposition results as centralized CPQR and establish an upper bound of the reconstruction error of FedGCUR. Extensive empirical results show that the proposed framework achieves higher accuracy compared to the baselines of data and feature selection methods, demonstrating its effectiveness and efficiency.

General Machine Learning · Causality

Xiu-Chuan Li, James Kwok, Jiaxian Guo, Tongliang Liu

It is a fundamental challenge to ascertain whether the causal effect of a treatment on an outcome is identifiable in the presence of latent confounders, which serves as the logical prerequisite for recovering the causal effect in a partially observed system. While prior literature demonstrates that the causal effect is identifiable when there exist auxiliary variables subject to stringent structural constraints, this paper investigates identifiability of the causal effect without such variables. This means that we ground identifiability solely in the joint distribution of the treatment-outcome pair, which constitutes the irreducible statistical basis for causal effect identification. Focusing on linear structural causal models (SCMs), we provide a nuanced and complete characterization of identifiability of the causal effect contingent on the distributional properties of exogenous noises. Specifically, we formulate a set of mutually exclusive and collectively exhaustive conditions regarding the Gaussianity of exogenous noises, ascertain under which conditions the causal effect is identifiable and under which it is not, while also quantifying the cardinality of the feasible solution set for the unidentifiable cases. Finally, we empirically validate our theoretical findings.

Deep Learning · Large Language Models

Egor Lifar, Semyon Savkin, Or Ordentlich, Yury Polyanskiy

This paper considers the problem of converting a given dense linear layer into a low-precision version. The tradeoff between minimizing description length and discrepancy introduced at the output of the layer is analyzed information theoretically (IT). It is shown that the popular GPTQ algorithm may have an arbitrarily large gap to IT limit. To alleviate this problem a novel algorithm, termed ''WaterSIC'', is proposed and is shown to be within a rate gap of 0.255 bit to IT limit, uniformly over all possible covariance matrices of input activations. WaterSIC's key innovation is allocating different quantization rates to different columns (in-features) of the weight matrix, mimicking the classical IT solution known as ''waterfilling''. Applying WaterSIC to real LLMs establishes new state-of-the-art for rates in the range of 1...4 bits per entry.

Applications · Robotics

Sanghwan Jang, Minjin Jeon, Minsoo Kim, Seong Jin Choi, Dongha Kim, Hwanjo Yu

Vision-Language-Action (VLA) models provide a versatile foundation for general robotic manipulation, yet they exhibit significant brittleness when confronted with novel task distributions. While In-Context Imitation Learning (ICIL) offers a training-free alternative, existing frameworks suffer from an *adaptation bottleneck* that hinders the effective translation of expert context to actions. This failure originates from superficial retrieval mechanisms and an entrenched *behavioral inertia* that anchors the policy to its pre-trained priors. To address these limitations, we present RA-VLA, a retrieval-augmented VLA framework that unifies behavior-aligned context retrieval with a grounded execution pipeline. By enforcing strict adherence to functional cues within a scalable architecture, our framework facilitates seamless task adaptation while preserving inference efficiency. Evaluations across the LIBERO benchmark and a real-world UR5e environment demonstrate that RA-VLA achieves superior success rates and computational efficiency, providing a robust framework for training-free robotic adaptation.

Applications · Time Series

Yangyou Liu, Zezhi Shao, Xinyu Chen, Hu Chen, Fei Wang, Yuankai Wu

Time series forecasting under non-stationarity faces a fundamental tension between capturing stable representations and adapting to distribution shifts. Existing methods implicitly rely on static historical assumptions, leading to a critical failure mode we term Phase Amnesia, where models become blind to the evolving global context. To resolve this, we formalize non-stationary dynamics through three physical hypotheses: Wold decomposition, dynamical phase evolution, and heteroscedastic manifold generation. These principles inspire PULSE, a physics-informed, plug-and-play framework adopting a Disentangle--Evolve--Simulate design philosophy. Specifically, PULSE utilizes phase-anchored disentanglement to resolve optimization interference caused by dominant trends, employs a Phase Router to actively generate future trajectories, and introduces Statistic-Aware Mixup (SAM) to ensure robustness against out-of-distribution volatility. Empirically, PULSE enables a simple MLP backbone to consistently outperform state-of-the-art Transformers across 12 real-world benchmarks. This validates that a correct physics-informed inductive bias is far more critical than raw architectural complexity for non-stationary forecasting.

Reinforcement Learning · Everything Else

Bosun Liang, Shuo Pei, Zirui Chen, Chuanzhi Fan, Chen Sun, Yuankai Wu, Huachun Tan, Yong Wang

Reinforcement learning often produces high-frequency oscillatory control signals that undermine the safety and stability required for physical deployment. Explicit action chunking addresses this by predicting fixed-horizon trajectories but increases the policy output dimension to R^hd, leading to optimization difficulties and incompatibility with standard step-wise interaction. To overcome these challenges, this paper proposes Dual-Window Smoothing (DWS), an implicit action chunking framework for smooth continuous control. Unlike explicit methods, DWS enforces temporal coherence without expanding the action space. It uses a dual-window design: an execution window that ensures physical smoothness through deterministic modulation, and a value window that aligns temporal-difference targets over the horizon to correct critic bias caused by open-loop execution. DWS also includes a lightweight actor-side temporal regularizer based on first-order action differences to promote global continuity. This design effectively bridges the gap between temporal abstraction and reactive step-wise control. Experiments on benchmarks including the DeepMind Control Suite and industrial energy management tasks with vector states show that DWS outperforms state-of-the-art (SOTA) baselines such as LipsNet++ and SmODE. In complex vision-based autonomous driving tasks, DWS achieves smoother control, safer behavior with reduced jitter, and attains a 100% success rate.

Deep Learning · Large Language Models

Jaewoo Jeong, Taesoo Kim, Sangdon Park

The hallucination of code generation models hinders their applicability to systems requiring higher safety standards. One critical bottleneck in addressing code hallucination is the difficulty of identifying the functional correctness of generated code, due to its unnatural form. We address this core bottleneck by automatically generating unit tests using dynamic code analysis tools, leveraging the \emph{executable nature} of code. Accordingly, we propose \emph{selective code generator} that abstains from uncertain generations -- based on the functional correctness evaluated by generated unit tests -- to theoretically control the correctness among non-abstained answers, \ie the false discovery rate. Finally, we propose to use generated unit tests in evaluation as well as in learning for precise code evaluation, calling this paradigm \emph{FuzzEval}. We demonstrate the efficacy of our method along with the controllability of code hallucination and reasonable selection efficiency.

Deep Learning · Generative Models and Autoencoders

Jiaoyang Ruan, Xin Gao, Yinda Chen, Hengyu Zeng, Liang Du, Guanghao Li, Jie Fu, Jian Pu

While Diffusion Large Language Models (dLLMs) offer structural advantages for global planning, efficiently verifying that they arrive at correct answers via valid reasoning traces remains a critical challenge. In this work, we propose a geometric perspective: Reasoning on the Manifold. We hypothesize that valid generation trajectories reside as stable attractors on the high-density manifold of the learned distribution, whereas invalid paths exhibit off-manifold drift. To operationalize this, we introduce Bidirectional Manifold Consistency (BMC), a training-free, unsupervised metric that quantifies the stability of the generated sequence through a forward-masking and backward-reconstruction cycle. Empirically, we demonstrate BMC's versatility across the full reasoning lifecycle: (1) in Diagnosis, it serves as a robust discriminator of solution validity without ground truth answer; (2) in Inference, it enables rejection resampling to effectively concentrate computational resources on complex reasoning tasks; and (3) in Alignment, it functions as a dense geometric reward that transforms sparse outcome supervision into fine-grained guidance, empowering models to self-evolve beyond standard baselines. Our results establish intrinsic geometric stability as a robust indicator of correctness for dLLMs.

Deep Learning · Generative Models and Autoencoders

Haoyang Chen, Jing Zhang, Hebaixu Wang, Shiqin Wang, Pohsun Huang, Jiayuan Li, Haonan Guo, Di Wang, Zheng Wang, Bo Du

Multi-modal remote sensing imagery provides complementary observations of the same geographic scene, yet such observations are frequently incomplete in practice. Existing cross-modal translation methods treat each modality pair as an independent task, resulting in quadratic complexity and limited generalization to unseen modality combinations. We formulate Any-to-Any translation as inference over a shared latent representation of the scene, where different modalities correspond to partial observations of the same underlying semantics. Based on this formulation, we propose Any2Any, a unified latent diffusion framework that projects heterogeneous inputs into a geometrically aligned latent space. Such structure performs anchored latent regression with a shared backbone, decoupling modality-specific representation learning from semantic mapping. Moreover, lightweight target-specific residual adapters are used to correct systematic latent mismatches without increasing inference complexity. To support learning under sparse but connected supervision, we introduce RST-1M, the first million-scale remote sensing dataset with paired observations across five sensing modalities, providing supervision anchors for any-to-any translation. Experiments across 14 translation tasks show that Any2Any consistently outperforms pairwise translation methods and exhibits strong zero-shot generalization to unseen modality pairs.

Deep Learning · Foundation Models

Yijie Tong, Yifan Hou, Shaobo Cui, Antoine Bosselut, Mrinmaya Sachan

Test-time compute (TTC) strategies have emerged as a lightweight approach to boost reasoning in large language models, but their applicability to vision-language models (VLMs) remains unclear. We present a systematic study of TTC for visual reasoning across seven open-source VLMs and six benchmarks, revisiting two paradigms: (i) feature-based scoring of chain-of-thought (CoT) traces and (ii) confidence-based aggregation via majority voting (MV). In the single-model setting, feature cues (e.g., length, pivot words) fail to improve accuracy, while MV yields only modest, CoT-dependent gains. To explain this limitation, we theoretically show that the voting method’s effectiveness depends on prediction diversity: when outputs are highly correlated, the benefit of voting vanishes. In contrast, multi-model ensembles introduce stronger diversity through architectural differences, training data, and scale, making them both more realistic and more promising for TTC. However, MV treats all models equally, leaving it vulnerable to correlated errors from weaker models. To address this, we propose Entropy-based TTC, which selects the most confident prediction based on predictive entropy. Our method reduces to MV in the single-model case but, in ensembles, leverages confidence disparities to prioritize stronger models. We prove that our method theoretically outperforms MV under mild dependence assumptions, and empirically show that it consistently surpasses both MV and the best individual model across diverse visual reasoning benchmarks. This demonstrates that smaller models can enhance, rather than hinder, larger ones when combined appropriately, unlocking ensemble gains not achievable with existing TTC strategies.

Deep Learning · Generative Models and Autoencoders

Kijung Jeon, Michael Muehlebach, Molei Tao

Generative modeling within constrained sets is essential for scientific and engineering applications involving physical, geometric, or safety requirements (e.g., molecular generation, robotics). We present a unified framework for constrained diffusion models on generic nonconvex feasible sets $\Sigma$ that simultaneously enforces equality and inequality constraints throughout the diffusion process. Our framework incorporates both overdamped and underdamped dynamics for forward and backward sampling. A key algorithmic innovation is a computationally efficient landing mechanism that replaces costly and often ill-defined projections onto $\Sigma$, ensuring feasibility without iterative Newton solves or projection failures. By leveraging underdamped dynamics, we accelerate mixing toward the prior distribution, effectively alleviating the high simulation costs typically associated with constrained diffusion. Empirically, this approach reduces function evaluations and memory usage during both training and inference while preserving sample quality. On benchmarks featuring equality and mixed constraints, our method achieves comparable sample quality to state-of-the-art baselines while significantly reducing computational cost, providing a practical and scalable solution for diffusion on nonconvex feasible sets.

Deep Learning · Generative Models and Autoencoders

Yuxin Jiang, Yuchao Gu, Ivor Tsang, Mike Zheng Shou

Scaling action-controllable world models is limited by the scarcity of action labels. While latent action learning promises to extract control interfaces from unlabeled video, learned latents often fail to transfer across contexts: they entangle scene-specific cues and lack a shared coordinate system. This occurs because standard objectives operate only *within* each clip, providing no mechanism to align action semantics across contexts. Our key insight is that although actions are unobserved, their *semantic effects* are observable and can serve as a shared reference. We introduce **Seq$\Delta$-REPA**, a sequence-level control-effect alignment objective that anchors integrated latent action to temporal feature differences from a frozen, self-supervised video encoder. Building on this, we present **Olaf-World**, a pipeline that pretrains action-conditioned video world models from large-scale passive video. Extensive experiments demonstrate that our method learns a more structured latent action space, leading to stronger zero-shot action transfer and more data-efficient adaptation to new control interfaces than state-of-the-art baselines.

Deep Learning · Generative Models and Autoencoders

Jiangshan Wang, Zeqiang Lai, Jiarui Chen, Jiayi Guo, Hang Guo, Xiu Li, Xiangyu Yue, Chunchao Guo

Diffusion Transformers (DiT) have demonstrated remarkable generative capabilities but remain highly computationally expensive. Previous acceleration methods, such as pruning and distillation, typically rely on a fixed computational capacity, leading to insufficient acceleration and degraded generation quality. To address this limitation, we propose \textbf{Elastic Diffusion Transformer (E-DiT)}, an adaptive acceleration framework for DiT that effectively improves efficiency while maintaining generation quality. Specifically, we observe that the generative process of DiT exhibits substantial sparsity (i.e., some computations can be skipped with minimal impact on quality), and this sparsity varies significantly across samples. Motivated by this observation, E-DiT equips each DiT block with a lightweight router that dynamically identifies sample-dependent sparsity from the input latent. Each router adaptively determines whether the corresponding block can be skipped. If the block is not skipped, the router then predicts the optimal MLP width reduction ratio within the block. During inference, we further introduce a block-level feature caching mechanism that leverages router predictions to eliminate redundant computations in a training-free manner. Extensive experiments across 2D image (Qwen-Image and FLUX) and 3D asset (Hunyuan3D-3.0) demonstrate the effectiveness of E-DiT, achieving up to $\sim$2$\times$ speedup with negligible loss in generation quality.

Applications · Computer Vision

Kuan Zhang, Dongchen Liu, Qiyue Zhao, Jinkun Hou, Xinran Zhang, Qinlei Xie, Miao Liu, Yiming Li

Human gameplay is a visually grounded interaction loop in which players act, reflect on failures, and watch tutorials to refine strategies. Can Vision-Language Models (VLMs) also learn from video-based reflection? We present **GameVerse**, a comprehensive video game benchmark that enables a *reflective visual interaction loop*. Moving beyond traditional ***fire-and-forget*** evaluations, it uses a novel ***reflect-and-retry*** paradigm to assess how VLMs internalize visual experience and improve policies. To facilitate systematic and scalable evaluation, we also introduce a *cognitive hierarchical taxonomy* spanning 15 globally popular games, *dual action space* for both semantic and GUI control, and *milestone evaluation* using advanced VLMs to quantify progress. Our experiments show that VLMs benefit from video-based reflection in varied settings, and perform best by combining failure trajectories and expert tutorials—a *training-free* analogue to reinforcement learning (RL) plus supervised fine-tuning (SFT).

Deep Learning · Graph Neural Networks

Xiaodong He, Haolan He, Ruiyi Fang, Ming Sun, zhao kang

Graph foundation models (GFMs) seek transferable representations across graph domains but are limited by structural heterogeneity and incompatible node feature spaces. We propose Structure-Centric Graph Foundation Models (SCGFM), which treat graph topology as the primary source of transferable knowledge. Modeling graphs as metric measure spaces, SCGFM introduces learnable geometric bases that define a shared structural coordinate system. Graphs are aligned to these bases via Gromov–Wasserstein distances, yielding structure-aligned latent representations that accommodate heterogeneous graph topologies. To address feature incompatibility, SCGFM employs a structure-aware feature re-encoding mechanism that unifies node representations without assuming a fixed feature dimensionality or requiring dataset-specific preprocessing. Experiments on graph- and node-level tasks demonstrate strong in-domain and cross-domain generalization, outperforming existing GFM approaches.

Deep Learning · Generative Models and Autoencoders

Jaemoo Choi, Yuchen Zhu, Wei Guo, Petr Molodyk, Bo Yuan, Jinbin Bai, Yi Xin, Molei Tao, Yongxin Chen

Reinforcement learning has been widely applied to diffusion and flow models for visual tasks such as text-to-image generation. However, these tasks remain challenging because diffusion models have intractable likelihoods, which creates a barrier for directly applying popular policy-gradient type methods. Existing approaches primarily focus on crafting new objectives built on already heavily engineered LLM objectives, using ad hoc estimators for likelihood, without a thorough investigation into how such estimation affects overall algorithmic performance. In this work, we provide a systematic analysis of the RL design space by disentangling three factors: i) policy-gradient objectives, ii) likelihood estimators, and iii) rollout sampling schemes. We show that adopting an evidence lower bound (ELBO) based model likelihood estimator, computed only from the final generated sample, is the dominant factor enabling effective, efficient, and stable RL optimization, outweighing the impact of the specific policy-gradient loss functional. We validate our findings across multiple reward benchmarks using SD 3.5 Medium, and observe consistent trends across all tasks. Our method improves the GenEval score from $0.24$ to $0.95$ in $90$ GPU hours, which is $4.6\times$ more efficient than FlowGRPO and $2\times$ more efficient than the SOTA method DiffusionNFT without reward hacking.

Anton Lydike, Jun Bi, Jackson Woodruff

On-device AI has gained significant attention for enabling efficient, low-latency inference on edge devices. However, tight resource constraints on these platforms make the deployment of accurate and lightweight deep learning models challenging. In particular, advanced activation functions (AFs) like Swish and GELU often incur high inference overhead due to the lack of hardware fast-paths for exponentiation and division, restricting edge-ML applications to simple AFs like ReLU, limiting model accuracy. To address this, we propose FFCC, a compiler that automatically generates efficient approximations of AFs through floating-point reinterpretation. These functions don’t require hardware fast-paths meaning they remain fast on edge devices. They do not incur great accurate losses, and allowing use as post-training replacements without negatively impacting model final accuracy. FFCC takes a specification of AFs using basic floating-point operators and applies derivation rules to lower these expressions into efficient instruction sequences. Our experiments show that we can provide fast approximations of AFs, achieving order-of-magnitude speed ups over accurate baselines on Arm M7, delivering performance on-par with Hardswish, while beating it on accuracy. Additionally, we show that our approximations – unlike Hardswish – can be used as drop-in replacements of exact version post-training without loss of model accuracy.

Anton Lydike, Jun Bi, Jackson Woodruff

On-device AI has gained significant attention for enabling efficient, low-latency inference on edge devices. However, tight resource constraints on these platforms make the deployment of accurate and lightweight deep learning models challenging. In particular, advanced activation functions (AFs) like Swish and GELU often incur high inference overhead due to the lack of hardware fast-paths for exponentiation and division, restricting edge-ML applications to simple AFs like ReLU, limiting model accuracy. To address this, we propose FFCC, a compiler that automatically generates efficient approximations of AFs through floating-point reinterpretation. These functions don’t require hardware fast-paths meaning they remain fast on edge devices. They do not incur great accurate losses, and allowing use as post-training replacements without negatively impacting model final accuracy. FFCC takes a specification of AFs using basic floating-point operators and applies derivation rules to lower these expressions into efficient instruction sequences. Our experiments show that we can provide fast approximations of AFs, achieving order-of-magnitude speed ups over accurate baselines on Arm M7, delivering performance on-par with Hardswish, while beating it on accuracy. Additionally, we show that our approximations – unlike Hardswish – can be used as drop-in replacements of exact version post-training without loss of model accuracy.

Chandan Tankala, Krishna Balasubramanian

Dense associative memories (DAMs) store and retrieve patterns via energy-function based fixed points, but existing models are limited to vector representations. We extend DAMs to Gaussian densities equipped with the 2-Wasserstein distance. Our framework defines a log-sum-exp energy over stored distributions and a retrieval dynamics aggregating optimal transport maps in a Gibbs-weighted manner. Stationary points correspond to self-consistent Wasserstein barycenters, generalizing classical DAM fixed points. We prove exponential storage capacity and provide quantitative retrieval guarantees under Wasserstein perturbations. We validate the method on synthetic and real-world image (CelebA and CIFAR-10 datasets) and text (text8 and NLI corpus) datasets. By generalizing from vectors to distributions, our work bridges classical DAMs with modern generative modeling and paves way for distributional storage and retrieval in memory-augmented learning.