论文检索

Social Aspects · Security

Yuhao Wang, Shengfang ZHAI, Guanghao Jin, Yinpeng Dong, Linyi Yang, Jiaheng Zhang

Large Language Model (LLM)-based agents employ external and internal memory systems to handle complex, goal-oriented tasks, yet this exposes them to severe extraction attacks, and corresponding defenses are currently lacking. In this paper, we propose *MemPot*, the first theoretically verified defense framework against memory extraction attacks by injecting optimized honeypots into the memory. Through a two-stage optimization process, *MemPot* generates trap documents that maximize the retrieval probability for attackers while remaining inconspicuous to benign users. We model the detection process as Wald’s Sequential Probability Ratio Test (SPRT) and theoretically prove that *MemPot* achieves a lower average number of sampling rounds compared to optimal static detectors. Empirically, *MemPot* significantly outperforms state-of-the-art baselines, achieving a 50% improvement in detection AUROC and an 80% increase in True Positive Rate under low False Positive Rate constraints. Furthermore, our experiments confirm that *MemPot* incurs zero online inference latency and preserves the agent's utility on standard tasks, verifying its superiority in safety, harmlessness and efficiency.

Deep Learning · Large Language Models

Jinhao Duan, Zicheng Liu, Zijie Liu, Kaidi Xu, Tianlong Chen

Humans express uncertainty verbally via markers (e.g., "possible", "likely"), yet most LLM uncertainty quantification (UQ) relies on costing likelihood- or consistency-based signals. From a cognitive perspective, accurate verbal uncertainty reflects metacognitive monitoring, representing knowledge boundaries ("knowing that you don't know") to support regulation and information seeking. In this paper, we investigate: How LLMs diverge from humans in verbal uncertainty quantification? Can verbal markers reliably quantify LLM uncertainty? We curate a corpus of human uncertainty markers from psychology and decision-science literature and benchmark LLMs against it. We observe that LLMs encode verbal uncertainty with numerical levels that differ substantially from those of humans. We then introduce VOCAL, a novel optimization-based algorithm that learns an optimal uncertainty profile over uncertainty markers directly from LLM outputs. By fitting a marker–uncertainty mapping to best explain empirical correctness, VOCALdiscovers how much probability mass each verbal marker should convey, rather than estimating uncertainty via repeated sampling. VOCAL enables a direct, marker-level comparison of confidence semantics between humans and LLMs, disentangling mismatch and revealing systematic confidence disparities in verbal expressions.

Theory · Game Theory

Annalisa Barbara, Riccardo Poiani, Martino Bernasconi, Andrea Celli

We study a setting in which two players play a (possibly approximate) Nash equilibrium of a bimatrix game, while a learner observes only their actions and has no knowledge of the equilibrium or the underlying game. A natural question is whether the learner can rationalize the observed behavior by inferring the players' payoff functions. Rather than producing a single payoff estimate, inverse game theory aims to identify the entire set of payoffs consistent with observed behavior, enabling downstream use in, e.g., counterfactual analysis and mechanism design across applications like auctions, pricing, and security games. We focus on the problem of estimating the set of feasible payoffs with high probability and up to precision $\epsilon$ on the Hausdorff metric. We provide the first minimax-optimal rates for both exact and approximate equilibrium play, in zero-sum as well as general-sum games. Our results provide learning-theoretic foundations for set-valued payoff inference in multi-agent environments.

Deep Learning · Generative Models and Autoencoders

Thomas Sesmat, Gabriel Meseguer-Brocal, Geoffroy Peeters

Understanding what generative models retain from training data remains challenging, with implications for copyright and privacy. This question becomes particularly relevant as Rectified Flows power increasingly deployed systems. We analyze the interpolation path $X_\lambda = (1-\lambda)X_0 + \lambda X_1$ that defines Rectified Flow training. We show that train-test distinguishability follows a bell-shaped curve over $\lambda$, with a maximum whose location we derive in closed form under Gaussian assumptions. This signal accumulates during training while validation metrics remain stable. We validate these predictions on both audio and images, and show that the bell-shaped structure is universal while the peak prediction holds when our assumptions are satisfied. As a proof of concept, we implement a membership inference attack achieving 0.91 AUC by exploiting this $\lambda$-resolved structure.

General Machine Learning · Representation Learning

Damion Harvey, Steven James, Benjamin Rosman, Branden Ingram, Geraud Nangue Tasse

We consider the problem of unsupervised skill segmentation and hierarchical structure discovery in reinforcement learning. While recent approaches have sought to segment trajectories into reusable skills or options, most rely on action labels, rewards, or handcrafted annotations, limiting their applicability. We propose a method that segments unlabelled trajectories into skills and induces a hierarchical structure over them using a grammar-based approach. The resulting hierarchy captures both low-level behaviours and their composition into higher-level skills. We evaluate our approach in high-dimensional, pixel-based environments, including Craftax and the full, unmodified version of Minecraft. Using metrics for skill segmentation, reuse, and hierarchy quality, we find that our method consistently produces more structured and semantically meaningful hierarchies than existing baselines. Furthermore, as a proof of concept for utility, we demonstrate that these discovered hierarchies accelerate and stabilise learning on downstream reinforcement learning tasks.

Applications · Computer Vision

Shaoxiong Zhan, Yanlin Lai, Zheng Liu, Zijian Lin, Lin Hai, Xiaodong Cai, Shen Li, Wen Huang, Hai-Tao Zheng

Current Large Language Models have achieved Olympiad-level logic, yet Vision-Language Models paradoxically falter on elementary spatial tasks like block counting. This capability mismatch reveals a critical "spatial intelligence gap," where models fail to construct coherent 3D mental representations from 2D observations. We uncover this gap via diagnostic analyses showing the bottleneck is a missing view-consistent spatial interface rather than insufficient visual features or weak reasoning. To bridge this, we introduce **3ViewSense**, a framework that grounds spatial reasoning in Orthographic Views. Drawing on engineering cognition, we propose a "Simulate-and-Reason" mechanism that decomposes complex scenes into canonical orthographic projections to resolve geometric ambiguities. By aligning egocentric perceptions with these allocentric references, our method facilitates explicit mental rotation and reconstruction. Empirical results on spatial reasoning benchmarks demonstrate that our method significantly outperforms existing baselines, with consistent gains on occlusion-heavy counting and view-consistent spatial reasoning. The framework also improves the stability and consistency of spatial descriptions, offering a scalable path toward stronger spatial intelligence in multimodal systems.

Shaoxiong Zhan, Yanlin Lai, Zheng Liu, Zijian Lin, Lin Hai, Xiaodong Cai, Shen Li, Wen Huang, Hai-Tao Zheng

Current Large Language Models have achieved Olympiad-level logic, yet Vision-Language Models paradoxically falter on elementary spatial tasks like block counting. This capability mismatch reveals a critical "spatial intelligence gap," where models fail to construct coherent 3D mental representations from 2D observations. We uncover this gap via diagnostic analyses showing the bottleneck is a missing view-consistent spatial interface rather than insufficient visual features or weak reasoning. To bridge this, we introduce **3ViewSense**, a framework that grounds spatial reasoning in Orthographic Views. Drawing on engineering cognition, we propose a "Simulate-and-Reason" mechanism that decomposes complex scenes into canonical orthographic projections to resolve geometric ambiguities. By aligning egocentric perceptions with these allocentric references, our method facilitates explicit mental rotation and reconstruction. Empirical results on spatial reasoning benchmarks demonstrate that our method significantly outperforms existing baselines, with consistent gains on occlusion-heavy counting and view-consistent spatial reasoning. The framework also improves the stability and consistency of spatial descriptions, offering a scalable path toward stronger spatial intelligence in multimodal systems.

Shuaibiao Han, Ruiyang Ni, Zhiguo Yang, Changlong Li, Perley Xu, Wenjie Ruan

Foundation models, such as Diffusion Models (DMs) and Large Language Models (LLMs), are now widely integrated into digital systems. This widespread use introduces a specific security risk: generative backdoors. Unlike traditional models where backdoors cause simple classification errors, generative backdoors hide within the model’s output distribution. This makes them difficult to detect using standard pattern-based methods.This paper argues that current defensive strategies are insufficient for generative AI. \textbf{We propose Distributional Integrity, a framework that focuses on maintaining the stability and accuracy of the model's data distribution.} We identify two primary threats: backdoors within the model supply chain and the contamination of synthetic data pipelines. To address these, we advocate for a shift toward cross-modal certification and parameter-level verification. These methods aim to secure the AI-generated content (AIGC) ecosystem against inherited vulnerabilities.

Applications · Robotics

Thomas Tian, Yilin Wu, Andrea Bajcsy

This position paper argues that to obtain reliable embodied reward models, the community must invest in "bad" robot data: failed, suboptimal, error-prone, and even hazardous behaviors. While reward models are central to any foundation model's lifecycle, today's embodied reward models are trained primarily on successful behaviors. We analyze three state-of-the-art embodied reward models and find that they systematically over-reward behaviors that real human evaluators would penalize, including unsafe interactions, poor execution, and shortcut strategies that only superficially satisfy tasks. We attribute these failures to a key data gap: the scarcity of negative embodied data which is costly to collect and often filtered out or withheld in existing robotics datasets. Furthermore, we show that even modest exposure to real bad behavior data can improve alignment with human preferences and reduce costly false positives. We therefore call on the embodied AI community to curate and release their bad robot data, build synthetic bad data generation engines, develop more decentralized physical evaluation systems, and design benchmarks for fine-grained embodied reward model evaluations.

Deep Learning · Large Language Models

Wentao Qiu, Guanran Luo, Zhongquan Jian, Jingqi Gao, Meihong Wang, Qingqiang Wu

A central challenge in large-scale decision-making under incomplete information is estimating reliable probabilities. Recent approaches leverage Large Language Models (LLMs) to generate explanatory factors and elicit coarse-grained probability estimates. Typically, an LLM performs forward abduction to propose factors, each paired with two mutually exclusive attributes, and a Naïve Bayes model is trained over factor combinations to refine the final probabilities. However, the induced factor space is often sparse, leading to frequent ''unknown'' outcomes when the system cannot map a query context to any supported factor configuration. Simply expanding the factor set to increase coverage is ineffective: it amplifies statistical noise and introduces spurious correlations that violate the conditional-independence assumption, ultimately degrading stability and reliability.To address these limitations, we propose Anchor, an inference framework that orchestrates aggregated Bayesian inference over a hierarchically structured factor space. Anchor first constructs a dense and organized factor space via iterative generation and hierarchical clustering. It then performs context-aware mapping through hierarchical retrieval and refinement, substantially reducing ''unknown'' predictions. Finally, Anchor augments Naïve Bayes with a Causal Bayesian Network to capture latent dependencies among factors, relaxing the strict independence assumption. Experiments show that Anchor markedly reduces ''unknown'' predictions and produces more reliable probability estimates than direct LLM baselines, achieving state-of-the-art performance while significantly reducing time and token overhead.

Deep Learning · Algorithms

Huijie Ma, Haoyuan Xin, Lei Meng, Guanzhou Ke, Yongyong Chen, Guoqing Chao

In recent years, multi-view outlier detection (MVOD) has gained increasing attention, with the primary objective of recovering the underlying structure of normal data from outlier-contaminated multi-view datasets. However, this objective is hindered by two fundamental challenges:(i) outlier propagation, (ii) scale discrepancy. To address these issues, we propose RNAMOD (Reliable Neighborhood-Aware Multi-View Outlier Detection), which introduces the concept of reliability and constructs a reliable neighborhood structure to avoid outlier propagation. We introduce a leave-one-out directional consensus mechanism to align cross-view neighborhood structures while preventing scale discrepancy by aligning geometric directions that remain invariant to scaling. Extensive experiments on six benchmark datasets demonstrate that RNAMOD consistently outperforms state-of-the-art methods.

General Machine Learning · Evaluation

Haoyu Zhao, Ziran Yang, Jiawei Li, Deyuan Mike He, Zenan Li, Chi Jin, Venugopal Veeravalli, Aarti Gupta, Sanjeev Arora

Vericoding refers to the generation of formally verified code from rigorous specifications. Recent AI models show promise in vericoding, but a unified methodology for cross-paradigm evaluation is lacking. Existing benchmarks test only an individual language/tool (e.g., Dafny, Verus, and Lean) and each covers very different tasks, so the performance numbers are not directly comparable. We address this gap with AlgoVeri, a benchmark that evaluates vericoding of $77$ classical algorithms in each of Dafny, Verus, and Lean. By enforcing identical functional contracts, AlgoVeri reveals critical capability gaps in current models. While frontier models achieve tractable success in Dafny ($40.3$\% for Gemini-3 Flash), where high-level abstractions and SMT automation simplify the workflow, performance collapses under the systems-level memory constraints of Verus ($24.7$\%) and the explicit proof construction required by Lean (7.8\%). Beyond aggregate metrics, we uncover a sharp divergence in test-time compute dynamics: Gemini-3 effectively utilizes iterative repair to boost performance (e.g., tripling pass rates in Dafny), whereas GPT-OSS saturates early. Finally, our error analysis shows that language design affects the refinement trajectory: while Dafny allows models to focus on logical correctness, Verus and Lean trap models in persistent syntactic and semantic barriers.

General Machine Learning · Unsupervised and Semi-supervised Learning

Boyang Dai, Chaoqi Chen, Yizhou Yu

Generalized Category Discovery (GCD) aims to recognize known classes while autonomously discovering novel ones in open-world settings. However, current approaches primarily focus on designing clustering objectives, often overlooking a critical bottleneck: standard vision backbones yield high-rank, entangled token representations that are ill-suited for unsupervised discovery of latent concepts and structures. In this paper, we propose Compositional Primitive Fields (CPF-GCD), a novel representation learning framework that reshapes the feature space to make such latent structure identifiable by enforcing a low-rank compositional organization. Our core hypothesis is that all categories (known/novel) reside on a latent manifold spanned by a finite set of learnable visual primitives that capture reusable concepts. CPF instantiates this geometric constraint via a spatial field mechanism. Inserted between the backbone and the head, it projects noisy patch tokens onto this structured manifold, effectively decomposing images into reusable atomic parts and their spatial layouts. By explicitly modeling the spatial distribution of primitives, CPF enables novel categories to emerge naturally as new activation patterns over a shared vocabulary. Extensive experiments demonstrate that CPF serves as a generic, plug-and-play module that consistently boosts performance across diverse GCD baselines, validating that identifying and leveraging low-rank compositional structure is a crucial inductive bias for open-world recognition.

General Machine Learning · Methodology

Paras Chopra

This position paper argues that answers to metaphysical puzzles in AI (such as ``Can LLMs be conscious?'' or ''What is AGI?'') should be judged by their practical consequences rather than their supposed truth. Our key position is that metaphysical concepts earn their value through the new research directions they open. Drawing on Pragmatism, we propose a two-step framework–*productive confusion*–to navigate conceptual confusions: first, clarify the different meanings a metaphysical concept has in ordinary language, then use this understanding to invent new empirical research programs. We illustrate our framework with numerous examples and show how it inspires progress for cutting-edge AI research. We contrast our position with Scientific Realism (which supposes science reveals ultimate truths) and Quietism (which brushes aside metaphysical puzzles as useless). We end with a call to action that operationalizes our position for multiple stakeholders in the AI community including researchers, decision makers and reviewers.

General Machine Learning · Causality

Oana-Iuliana Popescu, Wiebke Günther, Martin Rabel, Jakob Runge

Many real-world systems exhibit both context- and time-dependent causal dynamics, where the dynamical system state also influences its context. For instance, soil moisture is driven by precipitation, yet also provides the context for heat-flux realization. We capture such dynamics in Structural Causal Models (SCMs) by introducing endogenous and time-dependent discrete context variables, also allowing for possibly lagged dependencies with the system variables. While context variables are discrete, they may also be proxies of continuous variables. The enabling assumptions for causal discovery of our model are either persistence of the context or sparsity of the context–system dependencies. We design two new PCMCI-based algorithms for causal discovery with endogenous context variables for time series and prove their soundness. A systematic evaluation on synthetic benchmarks and an application to a real-world land-atmosphere feedback problem demonstrate their effectiveness and applicability.

Deep Learning · Graph Neural Networks

Zhuo Xu, Lu Bai, Jincheng Li, Lixin Cui, Ming Li, Hangyuan Du, Yue Wang

Graph Neural Networks (GNNs) have achieved remarkable successes in graph analysis due to the Message-Passing (MP) mechanism, yet they struggle with heterophilic graphs where connected nodes often have distinct labels or dissimilar attributes. Graph Contrastive Learning (GCL) serves as a promising approach to extract the information beyond neighboring nodes, effectively mitigating the limitations of the MP mechanism in handling heterophilic graphs. Nevertheless, GCL faces two critical challenges when applied to heterophilic graphs, i.e., the potential distribution shift from data augmentation and the loss of robustness caused by high-frequency signals. To address these problems, we propose a novel model, namely the Spectral Fusion Contrastive Learning with Topology-Adaptive Graph Augmentation (SFCLTA) for unsupervised graph representation learning. Our method dynamically adjusts graph structures by a heterophily-aware augmentation strategy, and constrains high-frequency distortions by spectral regularization. We utilize the confidence-weighted fusion to enhance the robustness. Additionally, we introduce a feature reconstruction task as the prerequisites to explicitly mitigate feature-level distribution shifts. Experiments on multiple real-world datasets demonstrate that the proposed SFCLTA consistently outperforms baseline models in multiple tasks.

Deep Learning · Large Language Models

Haosen Ge, Hamsa Bastani, Osbert Bastani

Rapidly increasing AI capabilities have substantial real-world consequences, ranging from AI safety concerns to labor market consequences. The Model Evaluation & Threat Research (METR) report argues that AI capabilities have exhibited exponential growth since 2019. In this position paper, we argue that the data does not support exponential growth, even in shorter-term horizons. Whereas the METR study claims that fitting sigmoid/logistic curves results in inflection points far in the future, we fit a sigmoid curve to their current data and find that the inflection point has already passed. In addition, we propose a more complex model that decomposes AI capabilities into base capabilities and reasoning capabilities, exhibiting individual rates of improvement. We prove that this model supports our hypothesis that AI capabilities will exhibit an inflection point in the near future. Our goal is not to establish a rigorous forecast of our own, but to highlight the fragility of existing forecasts of exponential growth. Finally, we call for the design of more rigorous evaluation methodologies for AI forecasts, and for better academic discussion on this topic.

Maxime Meyer, Mario Michelessa, Caroline Chaux, Vincent Tan

We study the fundamental expressivity limits of transformer models by formalizing the notion of accessible sequences---those that a transformer can produce for some prompt---and characterizing how accessibility depends on prompt length and model parameters. Our analysis provides a theoretical explanation for previously observed empirical failures of transformers on simple sequence tasks---such as copying and cramming---and yields both qualitative and quantitative predictions that hold across a wide range of architectures and model sizes. We prove that (i) the maximal length of accessible sequences grows linearly with the prompt length, (ii) beyond a critical threshold the proportion of accessible sequences decays exponentially with sequence length, and (iii) the linear coefficient relating prompt length to accessible sequence length admits a theoretical upper bound. Notably, these results hold even with unbounded context and computation time. Experiments using a “cramming” procedure confirm the linear scaling, the post-threshold exponential decay, and the tightness of the theoretical upper bound on different sizes of Pythia, Llamma, and Qwen architectures.

Deep Learning · Large Language Models

Yanwei Yue, Guibin Zhang, Boci Peng, Xuanbo Fan, Jiaxin Guo, Qiankun Li, Yan Zhang

Memory agents, which depart from predefined memory-processing pipelines by endogenously managing the processing, storage, and retrieval of memories, have garnered increasing attention for their autonomy and adaptability. However, existing training paradigms remain constrained: agents often traverse long-horizon sequences of memory operations before receiving sparse and delayed rewards, which hinders truly end-to-end optimization of memory management policies. To address this limitation, we introduce Mem-T, an autonomous memory agent that interfaces with a lightweight hierarchical memory database to perform dynamic updates and multi-turn retrieval over streaming inputs. To effectively train long-horizon memory management capabilities, we further propose MoT-GRPO, a tree-guided reinforcement learning framework that transforms sparse terminal feedback into dense, step-wise supervision via memory operation tree backpropagation and hindsight credit assignment, thereby enabling the joint optimization of memory construction and retrieval. Extensive experiments demonstrate that Mem-T is \textbf{\ding{182} high-performing}, surpassing frameworks such as A-Mem and Mem0 by up to $14.94\\%$, and \textbf{\ding{183} economical}, operating on a favorable accuracy-efficiency Pareto frontier and reducing inference tokens per query by $\sim24.45\\%$ relative to GAM without sacrificing performance.

Social Aspects · Everything Else

Marco Bornstein, Amrit Singh Bedi

The race for artificial intelligence (AI) dominance often prioritizes scale over efficiency. Hyper-scaling is the common industry approach: larger models, more data, and as many computational resources as possible. Using more resources is a simpler path to improved AI performance. Thus, efficiency has been de-emphasized. Consequently, the need for costly computational resources has marginalized academics and smaller companies. Simultaneously, increased energy expenditure, due to growing AI use, has led to mounting environmental costs. In response to accessibility and sustainability concerns, this position paper argues for research into, and implementation of, market-based methods that incentivize AI efficiency. We believe that incentivizing efficient operations and approaches will reduce emissions while opening new opportunities for academics and smaller companies. As a call to action, we propose a cap-and-trade system for AI. Our system provably reduces computations for AI deployment, thereby lowering emissions and monetizing efficiency to the benefit of of academics and smaller companies.