Na Li

Hanjiang Hu, Yiyuan Pan, Jiaxing Li et al.

As AI systems increasingly assist humans in physical tasks, ensuring safety becomes paramount -- physical actions carry immediate and irreversible consequences that digital errors do not. We introduce the Vision-Language Embodied Safety Agent (VLESA), a framework that monitors human activities from egocentric video and triggers real-time safety interventions when dangerous actions are predicted. VLESA addresses intent-dependent safety where identical actions can be safe or dangerous depending on context. A dataset pairing egocentric frames with goal-conditioned safety annotations is introduced, enabling a goal-conditioned safety Q-filter trained via GRPO that evaluates actions with respect to inferred intent without retraining. On top of that, an intent-action prediction agent is proposed to jointly infer goals and predict future actions from video. On the ASIMOV-2.0 benchmark, VLESA achieves higher intervention accuracy at the exact ground-truth frame compared to baselines, while the GRPO-trained Q-filter improves action safety by over 41 percentage points through goal-conditioned constrained decoding. Code is available at https://github.com/HanjiangHu/VLESA.

Na Li, Chengda Wang, Mingju Gao et al.

Diffusion large language models (DLLMs) enable non-autoregressive generation by iteratively denoising corrupted token sequences with bidirectional context. Despite their ability to update multiple positions in parallel, inference remains costly due to the many denoising steps required for high-quality generation. We propose SAID, a Scaffold-Aware Iterative Decoding framework that accelerates DLLMs by reallocating computation across tokens. SAID first spends denoising computation on scaffold tokens to establish the coarse semantic structure, and then completes predictable detail tokens with fewer steps. We further adapt SAID to block-wise diffusion decoding and introduce Confidence-Hierarchical Layered Generation (CHLG), which assigns additional steps only to low-confidence tokens. Experiments on LLaDA-8B and LLaDA 1.5 across math, coding, and knowledge benchmarks show that SAID significantly accelerates DLLM inference with a maximum speedup of 9.1x while maintaining competitive performance. Our code is publicly available: https://github.com/TH-AI-Lab-PKU/SAID.

Runyu Zhang, Qinghua Liu, Huan Wang et al. · salesforce

This paper studies policy optimization algorithms for multi-agent reinforcement learning. We begin by proposing an algorithm framework for two-player zero-sum Markov Games in the full-information setting, where each iteration consists of a policy update step at each state using a certain matrix game algorithm, and a value update step with a certain learning rate. This framework unifies many existing and new policy optimization algorithms. We show that the state-wise average policy of this algorithm converges to an approximate Nash equilibrium (NE) of the game, as long as the matrix game algorithms achieve low weighted regret at each state, with respect to weights determined by the speed of the value updates. Next, we show that this framework instantiated with the Optimistic Follow-The-Regularized-Leader (OFTRL) algorithm at each state (and smooth value updates) can find an $\mathcal{\widetilde{O}}(T^{-5/6})$ approximate NE in $T$ iterations, and a similar algorithm with slightly modified value update rule achieves a faster $\mathcal{\widetilde{O}}(T^{-1})$ convergence rate. These improve over the current best $\mathcal{\widetilde{O}}(T^{-1/2})$ rate of symmetric policy optimization type algorithms. We also extend this algorithm to multi-player general-sum Markov Games and show an $\mathcal{\widetilde{O}}(T^{-3/4})$ convergence rate to Coarse Correlated Equilibria (CCE). Finally, we provide a numerical example to verify our theory and investigate the importance of smooth value updates, and find that using "eager" value updates instead (equivalent to the independent natural policy gradient algorithm) may significantly slow down the convergence, even on a simple game with $H=2$ layers.

Chaoyi Wu, Pengcheng Qiu, Jinxin Liu et al.

In this study, we present MedS-Bench, a comprehensive benchmark designed to evaluate the performance of large language models (LLMs) in clinical contexts. Unlike existing benchmarks that focus on multiple-choice question answering, MedS-Bench spans 11 high-level clinical tasks, including clinical report summarization, treatment recommendations, diagnosis, named entity recognition, and medical concept explanation, among others. We evaluated six leading LLMs, e.g., MEDITRON, Mistral, InternLM 2, Llama 3, GPT-4, and Claude-3.5 using few-shot prompting, and found that even the most sophisticated models struggle with these complex tasks. To address these limitations, we developed MedS-Ins, a large-scale instruction tuning dataset for medicine. MedS-Ins comprises 58 medically oriented language corpora, totaling 13.5 million samples across 122 tasks. To demonstrate the dataset's utility, we conducted a proof-of-concept experiment by performing instruction tuning on a lightweight, open-source medical language model. The resulting model, MMedIns-Llama 3, significantly outperformed existing models across nearly all clinical tasks. To promote further advancements in the application of LLMs to clinical challenges, we have made the MedS-Ins dataset fully accessible and invite the research community to contribute to its expansion.Additionally, we have launched a dynamic leaderboard for MedS-Bench, which we plan to regularly update the test set to track progress and enhance the adaptation of general LLMs to the medical domain. Leaderboard: https://henrychur.github.io/MedS-Bench/. Github: https://github.com/MAGIC-AI4Med/MedS-Ins.

Pengfei Jin, Na Li, Quanzheng Li

Generative models are increasingly used as priors for inverse problems, but their ability to produce realistic images creates a basic trust problem: a plausible reconstruction may be supported by the measurements, or it may be filled in by the prior along unobserved directions. This distinction is especially important in medical imaging, where acquisition operators are designed under scan-time, dose, and calibration constraints. We study generative inverse problems from a measurement-geometry perspective. The central question is whether a fixed measurement operator can distinguish nearby images that are plausible under the generative prior, and whether this relationship can guide better measurements. We introduce a local measurement-manifold compatibility measure that quantifies how well the operator observes prior-relevant tangent directions. Under local regularity assumptions, we prove that this quantity controls the stable part of the reconstruction error, while the generative prior controls off-manifold drift. This worst-direction certificate motivates practical fixed and sequential acquisition rules based on overall local volume preservation, including a posterior-cloud design that adapts measurements at test time without training a sampling policy. Across row-sampling, tomographic, and MR acquisition settings, the proposed scores predict failure modes, explain measurement-induced hallucinations, and guide better sampling. In fastMRI Cartesian sampling, posterior-cloud measurement design improves over strong non-learned ACS-preserving baselines, including variable-density and Poisson-like masks.

Linwei Zhu, Yun Zhang, Na Li et al.

In intra coding, Rate Distortion Optimization (RDO) is performed to achieve the optimal intra mode from a pre-defined candidate list. The optimal intra mode is also required to be encoded and transmitted to the decoder side besides the residual signal, where lots of coding bits are consumed. To further improve the performance of intra coding in Versatile Video Coding (VVC), an intelligent intra mode derivation method is proposed in this paper, termed as Deep Learning based Intra Mode Derivation (DLIMD). In specific, the process of intra mode derivation is formulated as a multi-class classification task, which aims to skip the module of intra mode signaling for coding bits reduction. The architecture of DLIMD is developed to adapt to different quantization parameter settings and variable coding blocks including non-square ones, which are handled by one single trained model. Different from the existing deep learning based classification problems, the hand-crafted features are also fed into the intra mode derivation network besides the learned features from feature learning network. To compete with traditional method, one additional binary flag is utilized in the video codec to indicate the selected scheme with RDO. Extensive experimental results reveal that the proposed method can achieve 2.28%, 1.74%, and 2.18% bit rate reduction on average for Y, U, and V components on the platform of VVC test model, which outperforms the state-of-the-art works.

Bin Hu, Kaiqing Zhang, Na Li et al.

Gradient-based methods have been widely used for system design and optimization in diverse application domains. Recently, there has been a renewed interest in studying theoretical properties of these methods in the context of control and reinforcement learning. This article surveys some of the recent developments on policy optimization, a gradient-based iterative approach for feedback control synthesis, popularized by successes of reinforcement learning. We take an interdisciplinary perspective in our exposition that connects control theory, reinforcement learning, and large-scale optimization. We review a number of recently-developed theoretical results on the optimization landscape, global convergence, and sample complexity of gradient-based methods for various continuous control problems such as the linear quadratic regulator (LQR), $\mathcal{H}_\infty$ control, risk-sensitive control, linear quadratic Gaussian (LQG) control, and output feedback synthesis. In conjunction with these optimization results, we also discuss how direct policy optimization handles stability and robustness concerns in learning-based control, two main desiderata in control engineering. We conclude the survey by pointing out several challenges and opportunities at the intersection of learning and control.

Aoxiao Zhong, Hao He, Zhaolin Ren et al.

Cross-silo Federated learning (FL) has become a promising tool in machine learning applications for healthcare. It allows hospitals/institutions to train models with sufficient data while the data is kept private. To make sure the FL model is robust when facing heterogeneous data among FL clients, most efforts focus on personalizing models for clients. However, the latent relationships between clients' data are ignored. In this work, we focus on a special non-iid FL problem, called Domain-mixed FL, where each client's data distribution is assumed to be a mixture of several predefined domains. Recognizing the diversity of domains and the similarity within domains, we propose a novel method, FedDAR, which learns a domain shared representation and domain-wise personalized prediction heads in a decoupled manner. For simplified linear regression settings, we have theoretically proved that FedDAR enjoys a linear convergence rate. For general settings, we have performed intensive empirical studies on both synthetic and real-world medical datasets which demonstrate its superiority over prior FL methods.

Tongzheng Ren, Chenjun Xiao, Tianjun Zhang et al.

Deep latent variable models have achieved significant empirical successes in model-based reinforcement learning (RL) due to their expressiveness in modeling complex transition dynamics. On the other hand, it remains unclear theoretically and empirically how latent variable models may facilitate learning, planning, and exploration to improve the sample efficiency of RL. In this paper, we provide a representation view of the latent variable models for state-action value functions, which allows both tractable variational learning algorithm and effective implementation of the optimism/pessimism principle in the face of uncertainty for exploration. In particular, we propose a computationally efficient planning algorithm with UCB exploration by incorporating kernel embeddings of latent variable models. Theoretically, we establish the sample complexity of the proposed approach in the online and offline settings. Empirically, we demonstrate superior performance over current state-of-the-art algorithms across various benchmarks.

Runyu Zhang, Yang Hu, Na Li

Robust Markov Decision Processes (MDPs) and risk-sensitive MDPs are both powerful tools for making decisions in the presence of uncertainties. Previous efforts have aimed to establish their connections, revealing equivalences in specific formulations. This paper introduces a new formulation for risk-sensitive MDPs, which assesses risk in a slightly different manner compared to the classical Markov risk measure (Ruszczyński 2010), and establishes its equivalence with a class of soft robust MDP (RMDP) problems, including the standard RMDP as a special case. Leveraging this equivalence, we further derive the policy gradient theorem for both problems, proving gradient domination and global convergence of the exact policy gradient method under the tabular setting with direct parameterization. This forms a sharp contrast to the Markov risk measure, known to be potentially non-gradient-dominant (Huang et al. 2021). We also propose a sample-based offline learning algorithm, namely the robust fitted-Z iteration (RFZI), for a specific soft RMDP problem with a KL-divergence regularization term (or equivalently the risk-sensitive MDP with an entropy risk measure). We showcase its streamlined design and less stringent assumptions due to the equivalence and analyze its sample complexity

Philip K. Hopke, Maggie Leung, Na Li et al.

The ambient particulate chemical composition data with three particle diameter sizes (2.5mm<D< 1.15mm, 1.15mm<D<0.34mm and 0.34mm<D<0.1mm) collected at a major industrial center in Allen Park in Detroit, MI is examined. Standard multiway (tensor) methods like PARAFAC and Tucker tensor decompositions have been applied extensively to many chemical data. However, for multiple particle sizes, the source apportionment analysis calls for a novel multiway factor analysis. We apply the regularized block tensor decomposition to the collected air sample data. In particular, we use the Block Term Decomposition (BTD) in rank-(L;L;1) form to identify nine pollution sources (Fe+Zn, Sulfur with Dust, Road Dust, two types of Metal Works, Road Salt, Local Sulfate, and Homogeneous and Cloud Sulfate).

Na Li, Stefan Kindermann, Carmeliza Navasca

We study the convergence of the Regularized Alternating Least-Squares algorithm for tensor decompositions. As a main result, we have shown that given the existence of critical points of the Alternating Least-Squares method, the limit points of the converging subsequences of the RALS are the critical points of the least squares cost functional. Some numerical examples indicate a faster convergence rate for the RALS in comparison to the usual alternating least squares method.

Jiang Hu, Kangkang Deng, Na Li et al.

With a computationally efficient approximation of the second-order information, natural gradient methods have been successful in solving large-scale structured optimization problems. We study the natural gradient methods for the large-scale decentralized optimization problems on Riemannian manifolds, where the local objective function defined by the local dataset is of a log-probability type. By utilizing the structure of the Riemannian Fisher information matrix (RFIM), we present an efficient decentralized Riemannian natural gradient descent (DRNGD) method. To overcome the communication issue of the high-dimension RFIM, we consider a class of structured problems for which the RFIM can be approximated by a Kronecker product of two low-dimension matrices. By performing the communications over the Kronecker factors, a high-quality approximation of the RFIM can be obtained in a low cost. We prove that DRNGD converges to a stationary point with the best-known rate of $\mathcal{O}(1/K)$. Numerical experiments demonstrate the efficiency of our proposed method compared with the state-of-the-art ones. To the best of our knowledge, this is the first Riemannian second-order method for solving decentralized manifold optimization problems.

Hanjun Dai, Yuan Xue, Niao He et al.

In real-world decision-making, uncertainty is important yet difficult to handle. Stochastic dominance provides a theoretically sound approach for comparing uncertain quantities, but optimization with stochastic dominance constraints is often computationally expensive, which limits practical applicability. In this paper, we develop a simple yet efficient approach for the problem, the Light Stochastic Dominance Solver (light-SD), that leverages useful properties of the Lagrangian. We recast the inner optimization in the Lagrangian as a learning problem for surrogate approximation, which bypasses apparent intractability and leads to tractable updates or even closed-form solutions for gradient calculations. We prove convergence of the algorithm and test it empirically. The proposed light-SD demonstrates superior performance on several representative problems ranging from finance to supply chain management.

Yixu Huang, Bo Li, Na Li et al.

Generating a game is not the same as making one that can be played. Despite advances in code generation, existing approaches treat game generation as one-shot translation from prompt to artifact, leaving interaction-level failures undetected. We argue that evaluating and improving game generation requires a player, and study two roles for graphical user interface (GUI) agents in this process: (1) as an objective evaluator, for which we introduce PlaytestArena, a new evaluation environment that pairs 200 browser-based game generation tasks across eight genres with rubrics of expected in-play behaviors, adjudicated by a GUI agent that loads each build in a browser and plays it; and (2) as a subjective playtester, for which we propose Play2Code, where a game agent and a GUI agent operate in a sustained loop with shared memory, turning game generation into a dialogue between coding and playing. Our experiments show that even frontier models struggle to generate playable games directly, while Play2Code achieves a 66.8\% rubric pass-rate, improving over single-pass and agentic-coding baselines by 37.1 and 14.6 points respectively. Further analysis shows that GUI playtester feedback is more traceable than a human report, yet idiosyncratic in ways reminiscent of human testers, establishing game playtesting as a critical testbed for interactive code generation. Our project website is available at https://continual-game-generation.vercel.app/.

Xiuxian Li, Lihua Xie, Na Li

Distributed online optimization and game have been increasingly researched in the last decade, mostly motivated by its wide applications in sensor networks, robotics (e.g., distributed target tracking and formation control), smart grids, deep learning, and so forth. In these problems, there is a network of agents who may be cooperative (i.e., distributed online optimization) or noncooperative (i.e., online game) through local information exchanges. And the local cost function of each agent is often time-varying in dynamic and even adversarial environments. At each time, a decision must be made by each agent based on historical information at hand without knowing future information on cost functions. For these problems, a comprehensive survey is still lacking. This paper aims to provide a thorough overview of distributed online optimization and game from the perspective of problem settings, communication, computation, algorithms, and performances. In addition, some potential future directions are also discussed.

Haitong Ma, Tianpeng Zhang, Yixuan Wu et al.

We study the multi-agent Bayesian optimization (BO) problem, where multiple agents maximize a black-box function via iterative queries. We focus on Entropy Search (ES), a sample-efficient BO algorithm that selects queries to maximize the mutual information about the maximum of the black-box function. One of the main challenges of ES is that calculating the mutual information requires computationally-costly approximation techniques. For multi-agent BO problems, the computational cost of ES is exponential in the number of agents. To address this challenge, we propose the Gaussian Max-value Entropy Search, a multi-agent BO algorithm with favorable sample and computational efficiency. The key to our idea is to use a normal distribution to approximate the function maximum and calculate its mutual information accordingly. The resulting approximation allows queries to be cast as the solution of a closed-form optimization problem which, in turn, can be solved via a modified gradient ascent algorithm and scaled to a large number of agents. We demonstrate the effectiveness of Gaussian max-value Entropy Search through numerical experiments on standard test functions and real-robot experiments on the source-seeking problem. Results show that the proposed algorithm outperforms the multi-agent BO baselines in the numerical experiments and can stably seek the source with a limited number of noisy observations on real robots.

Amit Gajbhiye, Zied Bouraoui, Na Li et al.

Concepts play a central role in many applications. This includes settings where concepts have to be modelled in the absence of sentence context. Previous work has therefore focused on distilling decontextualised concept embeddings from language models. But concepts can be modelled from different perspectives, whereas concept embeddings typically mostly capture taxonomic structure. To address this issue, we propose a strategy for identifying what different concepts, from a potentially large concept vocabulary, have in common with others. We then represent concepts in terms of the properties they share with the other concepts. To demonstrate the practical usefulness of this way of modelling concepts, we consider the task of ultra-fine entity typing, which is a challenging multi-label classification problem. We show that by augmenting the label set with shared properties, we can improve the performance of the state-of-the-art models for this task.

Na Li

Motivated by the increasing need of saving search effort by obtaining relevant video clips instead of whole videos, we propose a new task, named Semantic Video Moments Retrieval at scale (SVMR), which aims at finding relevant videos coupled with re-localizing the video clips in them. Instead of a simple combination of video retrieval and video re-localization, our task is more challenging because of several essential aspects. In the 1st stage, our SVMR should take into account the fact that: 1) a positive candidate long video can contain plenty of irrelevant clips which are also semantically meaningful. 2) a long video can be positive to two totally different query clips if it contains clips relevant to two queries. The 2nd re-localization stage also exhibits different assumptions from existing video re-localization tasks, which hold an assumption that the reference video must contain semantically similar segments corresponding to the query clip. Instead, in our scenario, the retrieved long video can be a false positive one due to the inaccuracy of the first stage. To address these challenges, we propose our two-stage baseline solution of candidate videos retrieval followed by a novel attention-based query-reference semantically alignment framework to re-localize target clips from candidate videos. Furthermore, we build two more appropriate benchmark datasets from the off-the-shelf ActivityNet-1.3 and HACS for a thorough evaluation of SVMR models. Extensive experiments are carried out to show that our solution outperforms several reference solutions.

Michael Brazell, Na Li, Carmeliza Navasca et al.

Higher order tensor inversion is possible for even order. We have shown that a tensor group endowed with the Einstein (contracted) product is isomorphic to the general linear group of degree $n$. With the isomorphic group structures, we derived new tensor decompositions which we have shown to be related to the well-known canonical polyadic decomposition and multilinear SVD. Moreover, within this group structure framework, multilinear systems are derived, specifically, for solving high dimensional PDEs and large discrete quantum models. We also address multilinear systems which do not fit the framework in the least-squares sense, that is, when the tensor has an odd number of modes or when the tensor has distinct dimensions in each modes. With the notion of tensor inversion, multilinear systems are solvable. Numerically we solve multilinear systems using iterative techniques, namely biconjugate gradient and Jacobi methods in tensor format.

Chulong Zhang, Yuming Jiang, Na Li et al.

Pathologists need to combine information from differently stained pathology slices for accurate diagnosis. Deformable image registration is a necessary technique for fusing multi-modal pathology slices. This paper proposes a hybrid deep feature-based deformable image registration framework for stained pathology samples. We first extract dense feature points via the detector-based and detector-free deep learning feature networks and perform points matching. Then, to further reduce false matches, an outlier detection method combining the isolation forest statistical model and the local affine correction model is proposed. Finally, the interpolation method generates the deformable vector field for pathology image registration based on the above matching points. We evaluate our method on the dataset of the Non-rigid Histology Image Registration (ANHIR) challenge, which is co-organized with the IEEE ISBI 2019 conference. Our technique outperforms the traditional approaches by 17% with the Average-Average registration target error (rTRE) reaching 0.0034. The proposed method achieved state-of-the-art performance and ranked 1st in evaluating the test dataset. The proposed hybrid deep feature-based registration method can potentially become a reliable method for pathology image registration.

Bo Dai, Na Li, Dale Schuurmans

Self-supervised learning (SSL) have improved empirical performance by unleashing the power of unlabeled data for practical applications. Specifically, SSL extracts the representation from massive unlabeled data, which will be transferred to a plenty of down streaming tasks with limited data. The significant improvement on diverse applications of representation learning has attracted increasing attention, resulting in a variety of dramatically different self-supervised learning objectives for representation extraction, with an assortment of learning procedures, but the lack of a clear and unified understanding. Such an absence hampers the ongoing development of representation learning, leaving a theoretical understanding missing, principles for efficient algorithm design unclear, and the use of representation learning methods in practice unjustified. The urgency for a unified framework is further motivated by the rapid growth in representation learning methods. In this paper, we are therefore compelled to develop a principled foundation of representation learning. We first theoretically investigate the sufficiency of the representation from a spectral representation view, which reveals the spectral essence of the existing successful SSL algorithms and paves the path to a unified framework for understanding and analysis. Such a framework work also inspires the development of more efficient and easy-to-use representation learning algorithms with principled way in real-world applications.

Na Li, Hangguan Shan, Wei Ni et al.

Actor-critic (AC) methods are a cornerstone of reinforcement learning (RL) but offer limited interpretability. Current explainable RL methods seldom use state attributions to assist training. Rather, they treat all state features equally, thereby neglecting the heterogeneous impacts of individual state dimensions on the reward. We propose RKHS-SHAP-based Advanced Actor-Critic (RSA2C), an attribution-aware, kernelized, two-timescale AC algorithm, including Actor, Value Critic, and Advantage Critic. The Actor is instantiated in a vector-valued reproducing kernel Hilbert space (RKHS) with a Mahalanobis-weighted operator-valued kernel, while the Value Critic and Advantage Critic reside in scalar RKHSs. These RKHS-enhanced components use sparsified dictionaries: the Value Critic maintains its own dictionary, while the Actor and Advantage Critic share one. State attributions, computed from the Value Critic via RKHS-SHAP (kernel mean embedding for on-manifold and conditional mean embedding for off-manifold expectations), are converted into Mahalanobis-gated weights that modulate Actor gradients and Advantage Critic targets. We derive a global, non-asymptotic convergence bound under state perturbations, showing stability through the perturbation-error term and efficiency through the convergence-error term. Empirical results on three continuous-control environments show that RSA2C achieves efficiency, stability, and interpretability.

Na Li, John D. Kelleher, Robert Ross

In order to enhance levels of engagement with conversational systems, our long term research goal seeks to monitor the confusion state of a user and adapt dialogue policies in response to such user confusion states. To this end, in this paper, we present our initial research centred on a user-avatar dialogue scenario that we have developed to study the manifestation of confusion and in the long term its mitigation. We present a new definition of confusion that is particularly tailored to the requirements of intelligent conversational system development for task-oriented dialogue. We also present the details of our Wizard-of-Oz based data collection scenario wherein users interacted with a conversational avatar and were presented with stimuli that were in some cases designed to invoke a confused state in the user. Post study analysis of this data is also presented. Here, three pre-trained deep learning models were deployed to estimate base emotion, head pose and eye gaze. Despite a small pilot study group, our analysis demonstrates a significant relationship between these indicators and confusion states. We understand this as a useful step forward in the automated analysis of the pragmatics of dialogue.

Mingming Zhang, Feiqing Zhuang, Na Li et al.

Automated bidding is central to modern digital advertising. Early rule-based methods lacked adaptability, while subsequent Reinforcement Learning approaches modeled bidding as a Markov Decision Process but struggled with long-term dependencies. Recent generative models show promise, yet they lack explicit mechanisms to balance exploration and safety, relying solely on action perturbations or trajectory guidance without a safety fallback. This results in inefficient exploration and elevated financial risk for advertising platforms. To address this gap, we propose GUIDE (Generative Auto-Bidding with Unified Modeling and Exploration), a framework that synergistically integrates directed exploration with a safe fallback mechanism. GUIDE employs a Decision Transformer (DT) to jointly model historical bidding actions and environmental state transitions. A Q-value module guides the DT's exploration via regularization constraints, while an Inverse Dynamics Module (IDM) leverages DT-predicted future states to infer robust, behaviorally consistent actions as a safe policy fallback. The Q-value module then adaptively selects the final action between these two options, balancing exploration and safety. Together, these components form an integrated "explore-safeguard-select" pipeline that unifies efficiency and safety. We conduct extensive experiments on public datasets, in simulated auction environments, and through large-scale online deployment on Taobao, a leading Chinese advertising platform. Results show GUIDE consistently outperforms state-of-the-art baselines across all scenarios. In real-world deployment, GUIDE achieves notable gains: +4.10% ad GMV, +1.40% ad clicks, +1.66% ad cost, and +3.52% ad ROI, demonstrating its effectiveness and strong industrial applicability.

Xinhe Zhang, Yuyang Zhang, Pengfei Jin et al.

High-resolution 3D medical image generation remains challenging because fully volumetric models are computationally expensive, while efficient 2D slice generators often fail to preserve anatomical consistency across the third dimension. We propose LiFT, a framework for Lifted inter-slice Feature Trajectories that factorizes 3D volume synthesis into per-slice image generation and inter-slice trajectory learning. Rather than modeling the volumetric distribution end-to-end, LiFT treats a volume as an ordered trajectory in feature space, capturing how anatomical structures appear, transform, and disappear across depth. A tri-planar drifting loss aligns the trajectory of generated slices with the trajectories of real volumes, enabling distributional learning over inter-slice progressions in unconditional generation; in paired translation, a bidirectional $z$-context mixer trained against the registered target supplies through-plane coherence while preserving per-slice fidelity. We evaluate LiFT on BraTS 2023 (unconditional and missing-modality MR) and SynthRAD2023 (MR-to-CT). Across these settings, LiFT preserves per-slice quality, approaches the reported cWDM missing-MR reconstruction quality at $\sim$$135\times$ lower inference cost (without formal equivalence testing), and improves through-plane coherence on MR-to-CT relative to a no-mapper ablation, demonstrating that lightweight inter-slice trajectory learning is a viable route to high-resolution 3D medical synthesis.

Shen Li, Yuyang Zhang, Zhaolin Ren et al.

Interactive preference learning systems infer human preferences by presenting queries as pairs of options and collecting binary choices. Although binary choices are simple and widely used, they provide limited information about preference strength. To address this, we leverage human response times, which are inversely related to preference strength, as an additional signal. We propose a computationally efficient method that combines choices and response times to estimate human utility functions, grounded in the EZ diffusion model from psychology. Theoretical and empirical analyses show that for queries with strong preferences, response times complement choices by providing extra information about preference strength, leading to significantly improved utility estimation. We incorporate this estimator into preference-based linear bandits for fixed-budget best-arm identification. Simulations on three real-world datasets demonstrate that using response times significantly accelerates preference learning compared to choice-only approaches. Additional materials, such as code, slides, and talk video, are available at https://shenlirobot.github.io/pages/NeurIPS24.html

Zhaolin Ren, Tongzheng Ren, Haitong Ma et al.

This paper proposes an approach, Spectral Dynamics Embedding Control (SDEC), to optimal control for nonlinear stochastic systems. This method reveals an infinite-dimensional feature representation induced by the system's nonlinear stochastic dynamics, enabling a linear representation of the state-action value function. For practical implementation, this representation is approximated using finite-dimensional truncations, specifically via two prominent kernel approximation methods: random feature truncation and Nystrom approximation. To characterize the effectiveness of these approximations, we provide an in-depth theoretical analysis to characterize the approximation error arising from the finite-dimension truncation and statistical error due to finite-sample approximation in both policy evaluation and policy optimization. Empirically, our algorithm performs favorably against existing stochastic control algorithms on several benchmark problems.

Tianpeng Zhang, Kasper Johansson, Na Li

The multi-armed bandit(MAB) problem is a simple yet powerful framework that has been extensively studied in the context of decision-making under uncertainty. In many real-world applications, such as robotic applications, selecting an arm corresponds to a physical action that constrains the choices of the next available arms (actions). Motivated by this, we study an extension of MAB called the graph bandit, where an agent travels over a graph to maximize the reward collected from different nodes. The graph defines the agent's freedom in selecting the next available nodes at each step. We assume the graph structure is fully available, but the reward distributions are unknown. Built upon an offline graph-based planning algorithm and the principle of optimism, we design a learning algorithm, G-UCB, that balances long-term exploration-exploitation using the principle of optimism. We show that our proposed algorithm achieves $O(\sqrt{|S|T\log(T)}+D|S|\log T)$ learning regret, where $|S|$ is the number of nodes and $D$ is the diameter of the graph, which matches the theoretical lower bound $Ω(\sqrt{|S|T})$ up to logarithmic factors. To our knowledge, this result is among the first tight regret bounds in non-episodic, un-discounted learning problems with known deterministic transitions. Numerical experiments confirm that our algorithm outperforms several benchmarks.

Yuchen Jiao, Na Li, Changxiao Cai et al.

Higher-order ODE solvers have become a standard tool for accelerating diffusion probabilistic model (DPM) sampling, motivating the widespread view that first-order methods are inherently slower and that increasing discretization order is the primary path to faster generation. This paper challenges this belief and revisits acceleration from a complementary angle: beyond solver order, the placement of DPM evaluations along the reverse-time dynamics can substantially affect sampling accuracy in the low-neural function evaluation (NFE) regime. We propose a novel training-free, first-order sampler whose leading discretization error has the opposite sign to that of DDIM. Algorithmically, the method approximates the forward-value evaluation via a cheap one-step lookahead predictor. We provide theoretical guarantees showing that the resulting sampler provably approximates the ideal forward-value trajectory while retaining first-order convergence. Empirically, across standard image generation benchmarks (CIFAR-10, ImageNet, FFHQ, and LSUN), the proposed sampler consistently improves sample quality under the same NFE budget and can be competitive with, and sometimes outperform, state-of-the-art higher-order samplers. Overall, the results suggest that the placement of DPM evaluations provides an additional and largely independent design angle for accelerating diffusion sampling.

Xinyi Lin, Yuyang Zhang, Yuanhang Gan et al.

Scientific experiment and manufacture rely on complex, multi-step procedures that demand continuous human expertise for precise execution and decision-making. Despite advances in machine learning and automation, conventional models remain confined to virtual domains, while real-world experiment and manufacture still rely on human supervision and expertise. This gap between machine intelligence and physical execution limits reproducibility, scalability, and accessibility across scientific and manufacture workflows. Here, we introduce human-AI co-embodied intelligence, a new form of physical AI that unites human users, agentic AI, and wearable hardware into an integrated system for real-world experiment and intelligent manufacture. In this paradigm, humans provide precise execution and control, while agentic AI contributes memory, contextual reasoning, adaptive planning, and real-time feedback. The wearable interface continuously captures the experimental and manufacture processes, facilitates seamless communication between humans and AI for corrective guidance and interpretable collaboration. As a demonstration, we present Agentic-Physical Experimentation (APEX) system, coupling agentic reasoning with physical execution through mixed-reality. APEX observes and interprets human actions, aligns them with standard operating procedures, provides 3D visual guidance, and analyzes every step. Implemented in a cleanroom for flexible electronics fabrication, APEX system achieves context-aware reasoning with accuracy exceeding general multimodal large language models, corrects errors in real time, and transfers expertise to beginners. These results establish a new class of agentic-physical-human intelligence that extends agentic reasoning beyond computation into the physical domain, transforming scientific research and manufacturing into autonomous, traceable, interpretable, and scalable processes.

Yuyang Zhang, Derya Cansever, Na Li

This paper focuses on learning efficient sensor allocations that ensure observability of unknown high-dimensional linear systems using only a small number of sensors. Existing methods either require an impractically large number of sensors or assume access to an observable allocation in advance. We propose a two-stage framework that overcomes these limitations: first, a novel system identification algorithm integrates information from multiple trajectories, each observing different subsets of state coordinates; then, a classic sensor allocation method is adapted to operate on the learned system parameters. Our non-asymptotic guarantees show that the proposed approach learns a sensor allocation with a near-optimal number of sensors when sensors can be allocated on any state coordinate. We further extend the results to settings with inaccessible state coordinates that are unavailable for sensor allocation.

Yifan Wang, David Stevens, Pranay Shah et al.

The growing demand for AI training data has transformed data annotation into a global industry, but traditional approaches relying on human annotators are often time-consuming, labor-intensive, and prone to inconsistent quality. We propose the Model-in-the-Loop (MILO) framework, which integrates AI/ML models into the annotation process. Our research introduces a collaborative paradigm that leverages the strengths of both professional human annotators and large language models (LLMs). By employing LLMs as pre-annotation and real-time assistants, and judges on annotator responses, MILO enables effective interaction patterns between human annotators and LLMs. Three empirical studies on multimodal data annotation demonstrate MILO's efficacy in reducing handling time, improving data quality, and enhancing annotator experiences. We also introduce quality rubrics for flexible evaluation and fine-grained feedback on open-ended annotations. The MILO framework has implications for accelerating AI/ML development, reducing reliance on human annotation alone, and promoting better alignment between human and machine values.

Na Li, Robert Ross

Confusion is a mental state triggered by cognitive disequilibrium that can occur in many types of task-oriented interaction, including Human-Robot Interaction (HRI). People may become confused while interacting with robots due to communicative or even task-centred challenges. To build a smooth and engaging HRI, it is insufficient for an agent to simply detect confusion; instead, the system should aim to mitigate the situation. In light of this, in this paper, we present our approach to a linguistic design of dialogue policies to build a dialogue framework to alleviate interlocutor confusion. We also outline our sketch and discuss challenges with respect to its operationalisation.

Na Li, Robert Ross

Human-robot studies are expensive to conduct and difficult to control, and as such researchers sometimes turn to human-avatar interaction in the hope of faster and cheaper data collection that can be transferred to the robot domain. In terms of our work, we are particularly interested in the challenge of detecting and modelling user confusion in interaction, and as part of this research programme, we conducted situated dialogue studies to investigate users' reactions in confusing scenarios that we give in both physical and virtual environments. In this paper, we present a combined review of these studies and the results that we observed across these two embodiments. For the physical embodiment, we used a Pepper Robot, while for the virtual modality, we used a 3D avatar. Our study shows that despite attitudinal differences and technical control limitations, there were a number of similarities detected in user behaviour and self-reporting results across embodiment options. This work suggests that, while avatar interaction is no true substitute for robot interaction studies, sufficient care in study design may allow well executed human-avatar studies to supplement more challenging human-robot studies.

Binghang Lu, Runyu Zhang, Changhong Mou et al.

Physics-informed neural networks (PINNs) provide a flexible framework for solving forward and inverse problems governed by partial differential equations (PDEs), but standard PINN training typically relies on soft penalty formulations that combine PDE residuals, data mismatch, and initial/boundary conditions using manually chosen weights. This often leads to ill-conditioning, sensitivity to loss weights, and poor constraint satisfaction. In this work, we reformulate PINN training as an equality-constrained optimization problem and propose a novel Adaptive Momentum Feedback Linearization Optimization for Hard Constrained PINN (AdamFLIP). The key idea is to view the constraint residuals as the output of a controlled dynamical system and to compute the Lagrange multiplier as a feedback input that locally drives these residuals toward stable linear contraction dynamics. AdamFLIP then applies Adam-style first- and second-moment adaptation to the resulting feedback-linearized Lagrangian gradient, combining principled constraint handling with the scalability and robustness of adaptive neural-network optimization. We test AdamFLIP on a range of benchmark forward and inverse PDE problem, and it consistently outperforms both the standard soft-constrained PINN and state-of-the-art constrained optimizers. Specifically, on the Navier--Stokes equations benchmark, AdamFLIP \textbf{reduces relative $L_2$ error by more than two thirds} for the predicted solution compared to the next best method. Our AdamFLIP framework provides an effective and computationally scalable hard constraint optimization method for PINN training.

Jin Du, Xinhe Zhang, Hao Shen et al.

Lifelong learning in artificial intelligence (AI) aims to mimic the biological brain's ability to continuously learn and retain knowledge, yet it faces challenges such as catastrophic forgetting. Recent neuroscience research suggests that neural activity in biological systems undergoes representational drift, where neural responses evolve over time, even with consistent inputs and tasks. We hypothesize that representational drift can alleviate catastrophic forgetting in AI during new task acquisition. To test this, we introduce DriftNet, a network designed to constantly explore various local minima in the loss landscape while dynamically retrieving relevant tasks. This approach ensures efficient integration of new information and preserves existing knowledge. Experimental studies in image classification and natural language processing demonstrate that DriftNet outperforms existing models in lifelong learning. Importantly, DriftNet is scalable in handling a sequence of tasks such as sentiment analysis and question answering using large language models (LLMs) with billions of parameters on a single Nvidia A100 GPU. DriftNet efficiently updates LLMs using only new data, avoiding the need for full dataset retraining. Tested on GPT-2 and RoBERTa, DriftNet is a robust, cost-effective solution for lifelong learning in LLMs. This study not only advances AI systems to emulate biological learning, but also provides insights into the adaptive mechanisms of biological neural systems, deepening our understanding of lifelong learning in nature.

Na Li, Binod Koirala

As energy communities move from policy design to implementation in Switzerland, understanding their performance in practice has become increasingly important. A techno-economic assessment of a regulation-compliant LEC is presented under the new Swiss legal framework in this study. A reference case without local electricity exchange is compared to a LEC scenario with internal electricity sharing. Results show that LEC participation increases local renewable utilization, reduces grid exports, and delivers economic benefits to both consumers and prosumers. A sensitivity analysis further indicates that internal electricity pricing plays a critical role in shaping trade-offs between overall efficiency and fairness in benefit distribution. This exploratory study provides practical insights to support informed decision-making and the future development of LEC in Switzerland.

Haitong Ma, Chenxiao Gao, Tianyi Chen et al.

A commonly used family of RL algorithms for diffusion policies conducts softmax reweighting over the behavior policy, which usually induces an over-greedy policy and fails to leverage feedback from negative samples. In this work, we introduce Signed Measure Policy Optimization (SiMPO), a simple and unified framework that generalizes reweighting scheme in diffusion RL with general monotonic functions. SiMPO revisits diffusion RL via a two-stage measure matching lens. First, we construct a virtual target policy by $f$-divergence regularized policy optimization, where we can relax the non-negativity constraint to allow for a signed target measure. Second, we use this signed measure to guide diffusion or flow models through reweighted matching. This formulation offers two key advantages: a) it generalizes to arbitrary monotonically increasing weighting functions; and b) it provides a principled justification and practical guidance for negative reweighting. Furthermore, we provide geometric interpretations to illustrate how negative reweighting actively repels the policy from suboptimal actions. Extensive empirical evaluations demonstrate that SiMPO achieves superior performance by leveraging these flexible weighting schemes, and we provide practical guidelines for selecting reweighting methods tailored to the reward landscape.

Yiqi Tian, Pengfei Jin, Mingze Yuan et al.

Diffusion models have achieved state-of-the-art performance in generative modeling, yet their sampling procedures remain vulnerable to hallucinations-often stemming from inaccuracies in score approximation. In this work, we reinterpret diffusion sampling through the lens of optimization and introduce RODS (Robust Optimization-inspired Diffusion Sampler), a novel method that detects and corrects high-risk sampling steps using geometric cues from the loss landscape. RODS enforces smoother sampling trajectories and adaptively adjusts perturbations, reducing hallucinations without retraining and at minimal additional inference cost. Experiments on AFHQv2, FFHQ, and 11k-hands demonstrate that RODS maintains comparable image quality and preserves generation diversity. More importantly, it improves both sampling fidelity and robustness, detecting over 70% of hallucinated samples and correcting more than 25%, all while avoiding the introduction of new artifacts. We release our code at https://github.com/Yiqi-Verna-Tian/RODS.

Yongchao Chen, Yueying Liu, Junwei Zhou et al.

Practical guidance on training Large Language Models (LLMs) to leverage Code Interpreter across diverse tasks remains lacking. We present R1-Code-Interpreter, an extension of a text-only LLM trained via multi-turn supervised fine-tuning (SFT) and reinforcement learning (RL) to autonomously generate multiple code queries during step-by-step reasoning. Unlike prior RL + tool-use efforts focused on narrow domains such as math or retrieval, we curate 144 diverse reasoning and planning tasks and show that training a general-purpose Code Interpreter across them presents significant challenges due to task heterogeneity and scarcity of effective samples. To address this, we introduce a multi-stage curriculum learning approach that partitions training samples by measured improvement potential. The RL training prioritizes samples with higher potential and gradually shifts to lower-potential ones, increasing the average RL gains from merely +3.4% to +9.3% across Qwen-2.5 models (3/7/14B). Our final model, R1-CI-14B, improves average accuracy on the 37 test tasks from 44.1% to 72.4%, outperforming text-only GPT-4o (58.6%) and GPT-4o with Code Interpreter (70.9%). Notably, R1-CI-14B also exhibits emergent self-checking behavior through code generation. Datasets, Codes, and Models are available at https://github.com/yongchao98/R1-Code-Interpreter and https://huggingface.co/yongchao98.

Na Li, Renyu Zhu, Xiaoxu Zhou et al.

Author disambiguation arises when different authors share the same name, which is a critical task in digital libraries, such as DBLP, CiteULike, CiteSeerX, etc. While the state-of-the-art methods have developed various paper embedding-based methods performing in a top-down manner, they primarily focus on the ego-network of a target name and overlook the low-quality collaborative relations existed in the ego-network. Thus, these methods can be suboptimal for disambiguating authors. In this paper, we model the author disambiguation as a collaboration network reconstruction problem, and propose an incremental and unsupervised author disambiguation method, namely IUAD, which performs in a bottom-up manner. Initially, we build a stable collaboration network based on stable collaborative relations. To further improve the recall, we build a probabilistic generative model to reconstruct the complete collaboration network. In addition, for newly published papers, we can incrementally judge who publish them via only computing the posterior probabilities. We have conducted extensive experiments on a large-scale DBLP dataset to evaluate IUAD. The experimental results demonstrate that IUAD not only achieves the promising performance, but also outperforms comparable baselines significantly. Codes are available at https://github.com/papergitgit/IUAD.

Chao Zhu, Teng Miao, Tongyu Xu et al.

Nowadays, there are many approaches to acquire three-dimensional (3D) point clouds of maize plants. However, automatic stem-leaf segmentation of maize shoots from three-dimensional (3D) point clouds remains challenging, especially for new emerging leaves that are very close and wrapped together during the seedling stage. To address this issue, we propose an automatic segmentation method consisting of three main steps: skeleton extraction, coarse segmentation based on the skeleton, fine segmentation based on stem-leaf classification. The segmentation method was tested on 30 maize seedlings and compared with manually obtained ground truth. The mean precision, mean recall, mean micro F1 score and mean over accuracy of our segmentation algorithm were 0.964, 0.966, 0.963 and 0.969. Using the segmentation results, two applications were also developed in this paper, including phenotypic trait extraction and skeleton optimization. Six phenotypic parameters can be accurately and automatically measured, including plant height, crown diameter, stem height and diameter, leaf width and length. Furthermore, the values of R2 for the six phenotypic traits were all above 0.94. The results indicated that the proposed algorithm could automatically and precisely segment not only the fully expanded leaves, but also the new leaves wrapped together and close together. The proposed approach may play an important role in further maize research and applications, such as genotype-to-phenotype study, geometric reconstruction and dynamic growth animation. We released the source code and test data at the web site https://github.com/syau-miao/seg4maize.git

Yuyang Zhang, Haldun Balim, Na Li

Cooperative multi-agent reinforcement learning (MARL) involves complex agent interactions and requires effective exploration strategies. A prominent class of MARL algorithms, decentralized softmax policy gradient (DecSPG), addresses this through energy-based policy updates. In practice, however, such energy-based policies are intractable to maintain and are commonly projected onto the Gaussian policy class. In this work, we show that the limited expressiveness of Gaussian policies severely hinders exploration in DecSPG, and this limitation worsens as the number of agents grows. To address this issue, we propose decentralized diffusion policy learning (DDPL), which parameterizes each agent's policy with a denoising diffusion probabilistic model, an expressive generative model that captures multi-modal action distributions for enhanced exploration. DDPL enables efficient online training of diffusion policies via importance sampling score matching (ISSM), a novel training method with theoretical guarantee. We evaluate DDPL on representative continuous-action MARL benchmarks, including multi-agent particle environment, multi-agent MuJoCo, IsaacLab, and JAX-reimplemented StarCraft multi-agent challenge, and observe consistently improved performance.

Na Li, Dong Li, Stavros Orfanoudakis et al.

The institutional separation between local energy communities and public electric vehicle (EV) charging limits the efficient use of locally generated renewable energy. This paper introduces the concept of community-to-vehicle (C2V) as an institutional design mechanism to bridge this gap by enabling EV charging within the community boundary, where locally generated photovoltaic (PV) surplus is preferentially allocated and offered to external users at a community charging price. Building on the recently introduced local electricity community framework in Switzerland, we design scenarios that capture the transition from full separation to coordinated EV charging and evaluate their impacts on EV users and the community. The results show that C2V significantly improves local PV utilization and enhances economic performance, reducing EV charging costs relative to commercial alternatives while generating additional revenue streams for the community. These findings highlight the potential of C2V as a practical, implementable mechanism for integrating EV charging into local energy communities, providing a clear pathway for adopting coordinated community-EV interaction within existing regulatory frameworks.

Haldun Balim, Na Li, Yilun Du

Offline decision-making requires synthesizing reliable behaviors from fixed datasets without further interaction, yet existing generative approaches often yield trajectories that are dynamically infeasible. We propose Model Predictive Diffuser (MPDiffuser), a compositional model-based diffusion framework consisting of: (i) a planner that generates diverse, task-aligned trajectories; (ii) a dynamics model that enforces consistency with the underlying system dynamics; and (iii) a ranker module that selects behaviors aligned with the task objectives. MPDiffuser employs an alternating diffusion sampling scheme, where planner and dynamics updates are interleaved to progressively refine trajectories for both task alignment and feasibility during the sampling process. We also provide a theoretical rationale for this procedure, showing how it balances fidelity to data priors with dynamics consistency. Empirically, the compositional design improves sample efficiency, as it leverages even low-quality data for dynamics learning and adapts seamlessly to novel dynamics. We evaluate MPDiffuser on both unconstrained (D4RL) and constrained (DSRL) offline decision-making benchmarks, demonstrating consistent gains over existing approaches. Furthermore, we present a preliminary study extending MPDiffuser to vision-based control tasks, showing its potential to scale to high-dimensional sensory inputs. Finally, we deploy our method on a real quadrupedal robot, showcasing its practicality for real-world control.

Zishuo Wan, Qinqin Kang, Na Li et al.

The accurate diagnosis and segmentation of tumors in contrast-enhanced Computed Tomography (CT) are fundamentally driven by the distinctive hemodynamic profiles of contrast agents over time. However, in real-world clinical practice, complete temporal dynamics are often hard to capture by strict radiation dose limits and inconsistent acquisition protocols across institutions, leading to a prevalent missing modality problem. Existing deep learning approaches typically treat missing phases as absent independent channels, ignoring the inherent temporal continuity of hemodynamics. In this work, we propose Time Attenuated Representation Disentanglement (TARDis), a novel physics-aware framework that redefines missing modalities as missing sample points on a continuous Time-Attenuation Curve. We first hypothesize that the latent feature can be disentangled into a time-invariant static component (anatomy) and a time-dependent dynamic component (perfusion). We achieve this via a dual-path architecture: a quantization-based path using a learnable embedding dictionary to extract consistent anatomical structures, and a probabilistic path using a Hemodynamic Conditional Variational Autoencoder to model dynamic enhancement conditioned on the estimated scan time. This design allows the network to infer missing hemodynamic features by sampling from the learned latent distribution. Extensive experiments on a large-scale multi-modal private abdominal CT dataset (2,282 patients) and two public datasets demonstrate that TARDis significantly outperforms state-of-the-art incomplete modality frameworks. Notably, our method maintains robust diagnostic performance even in extreme data-sparsity scenarios, highlighting its potential for reducing radiation exposure while maintaining diagnostic precision.

Yang Hu, Na Li

In this paper, we propose a novel framework for the joint identification of system dynamics and noise covariance in linear systems, under general noise distributions beyond Gaussian. Specifically, we would like to simultaneously estimate the dynamical matrix $A$ and the noise covariance matrix $\varSigma$ using state transition data. The formulation builds upon a novel parameterization of the state-transition distribution, which enables more effective use of distributional "shape" information for improved identification accuracy. We introduce two practical estimators, namely the maximum likelihood estimator (MLE) and the score-matching estimator (SME), to solve the joint dynamics-covariance identification problem, and provide rigorous analysis of their statistical properties and sample complexity. Simulation results show that the proposed estimators outperform the ordinary least squares (OLS) baseline.

Yuyang Zhang, Yang Hu, Bo Dai et al.

Soft actor-critic (SAC) is a popular algorithm for max-entropy reinforcement learning. In practice, the energy-based policies in SAC are often approximated using simple policy classes for efficiency, sacrificing the expressiveness and robustness. In this paper, we propose a variant of the SAC algorithm that parameterizes the policy with flow-based models, leveraging their rich expressiveness. In the algorithm, we evaluate the flow-based policy utilizing the instantaneous change-of-variable technique and update the policy with an online variant of flow matching developed in this paper. This online variant, termed importance sampling flow matching (ISFM), enables policy update with only samples from a user-specified sampling distribution rather than the unknown target distribution. We develop a theoretical analysis of ISFM, characterizing how different choices of sampling distributions affect the learning efficiency. Finally, we conduct a case study of our algorithm on the max-entropy linear quadratic regulator problems, demonstrating that the proposed algorithm learns the optimal action distribution.

Na Li, Chunyi Zhou, Yansong Gao et al.

Personal digital data is a critical asset, and governments worldwide have enforced laws and regulations to protect data privacy. Data users have been endowed with the right to be forgotten of their data. In the course of machine learning (ML), the forgotten right requires a model provider to delete user data and its subsequent impact on ML models upon user requests. Machine unlearning emerges to address this, which has garnered ever-increasing attention from both industry and academia. While the area has developed rapidly, there is a lack of comprehensive surveys to capture the latest advancements. Recognizing this shortage, we conduct an extensive exploration to map the landscape of machine unlearning including the (fine-grained) taxonomy of unlearning algorithms under centralized and distributed settings, debate on approximate unlearning, verification and evaluation metrics, challenges and solutions for unlearning under different applications, as well as attacks targeting machine unlearning. The survey concludes by outlining potential directions for future research, hoping to serve as a guide for interested scholars.