arXiv
2026-10-21
AI 研究进展 (cs.AI)
Existing sequential recommenders usually allocate a fixed computational budget to every item in a user's interaction history, regardless of the item's contribution to the final prediction. We challenge this de facto design with two diagnostic findings. (1) Non-uniform computational need: items withi…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Existing world models typically perceive a scene through current visual frame or a narrow history window. However, these observations may not reveal the physics properties that determine transition dynamics, such as an object's mass distribution, joint damping, or surface friction. Without calibrati…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Through a tangle of real-world interactions, from synthetic training pipelines to data shared on the open web, frontier LLMs absorb what their predecessors and competitors create. How well does each model play its role in this ecosystem, as a creator and as a solver? We introduce Environment Duels,…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Consensus is a fundamental problem in distributed computing and at the core of state machine replication (SMR). Despite decades of active human research, discovering practical low-latency solutions to the problem is still open, as it requires combining protocol knowledge, correctness reasoning, impl…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Quantitative research is an iterative loop: a researcher proposes a market hypothesis, implements it as an executable signal, tests it on historical data, and decides what to keep. Existing benchmarks score formulas, code, or individual trading decisions, leaving open whether large language models (…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Multimodal agents (MMAs) sustain interaction through reasoning, memory, tools, and feedback, most visibly as computer-use agents, while robotic systems (RSs) couple sensing and actuation under physical dynamics. We define multimodal embodied agents (MMEAs) as goal-directed systems that couple multim…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Sequential decision-making under partial observability is difficult because agents must explore to infer hidden, task-relevant information before acting effectively. Learning such information-gathering behaviors tabula rasa via online reinforcement learning is sample-inefficient, whereas offline imi…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Despite recent advances, current AI systems remain highly susceptible to domain shifts, particularly in biomedical imaging, where images are usually acquired under substantially different experimental conditions. Although recent domain adaptation methods have made substantial progress, significant p…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Reinforcement Learning with Human Feedback (RLHF) has become a central paradigm for aligning Large Language Models with human values. However, the reliance on sensitive preference data necessitates rigorous differential privacy (DP) guarantees to prevent user information leakage. Existing private al…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Vision-language-action (VLA) models have shown promise in robotic manipulation, but imitation learning (IL) alone often struggles to master complex long-horizon skills. Reinforcement learning (RL) offers an effective avenue for further improving VLA policies. However, existing VLA-RL methods typical…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Parallel AI agents can produce useful local edits yet degrade a shared solution when their edits are combined. We model this coordination problem as discrete block optimization with an exactly evaluable objective and a known dependency graph. Independent sets certify compatible edits whose verified…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Data agents built on ReAct-style loops now execute complex analytical workflows over structured data, yet their intermediate reasoning still runs almost entirely through text. Human analysts instead move between numbers, tables, and charts, reading visual evidence to expose trends and anomalies that…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Recent advances in frontier AI models have demonstrated remarkable capabilities in visual understanding, reasoning, planning, and tool use, raising the possibility of using general-purpose AI models for robot manipulation without task-specific training. However, their manipulation capabilities have…
arXiv
2026-10-21
AI 研究进展 (cs.AI)
Coding agents are typically evaluated using one-shot, well-formulated prompts. Benchmarks such as SWE-bench use existing GitHub issues as input for a coding agent and evaluate whether the agent successfully solves the task. However, in real-world use cases, prompts are often not as well-formulated o…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Reinforcement learning (RL) is an effective way to strengthen the long-context reasoning of large language models (LLMs), but it needs training data that are long, diverse, challenging, and verifiable. Human annotation of such data is expensive, naturally occurring data are too scarce to sustain RL,…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Existing sequential recommenders usually allocate a fixed computational budget to every item in a user's interaction history, regardless of the item's contribution to the final prediction. We challenge this de facto design with two diagnostic findings. (1) Non-uniform computational need: items withi…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Language models make it easy to build classifiers by simply prompting a model with the sample to classify. However, their predictions are neither deterministic nor uniformly reliable across samples. How can we combine predictions from different models to improve reliability? We view LLM classifiers…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Uniform diffusion language models (UDLMs) generate text by iteratively denoising sequences of ordinary vocabulary tokens, any of which may be corrupted and subsequently revised. During training, UDLMs must denoise from a context that may itself be unreliable, whereas masked diffusion language models…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Understanding large, complex, and rapidly evolving code bases remains a core challenge for software engineering at scale, hindering onboarding, maintenance, refactoring, and impact analysis. Providing this information to a Large-Language Model (LLM) is another hurdle, as the LLM must understand not…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Reinforcement learning (RL) based on PPO-style clipping has become an effective approach to improving the reasoning capabilities of large language models (LLMs). However, standard token-level clipping measures policy change only through the conditional probability of the current token, while the pre…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Benchmark performance is the dominant standard for comparing large language models (LLMs), typically summarized by a single aggregate score, such as accuracy, pass rate, or rubric score. This aggregate score provides limited information about a model's underlying capabilities, yet understanding thes…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Literature-searching language agents are becoming a standard component of scientific assistants for the life sciences, yet there is little controlled evidence on which parts of their harness matter. We run a controlled ablation on biomedical claim verification (SciFact) with the live PubMed API as t…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Chain-of-thought (CoT) monitoring is one of the few practical tools available for overseeing capable AI systems, but it is fragile: a model's reasoning is only useful to a monitor if it remains coupled to what the model actually does. This coupling can be degraded during training through, for exampl…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Using an LLM to generate training data for another is among the most general ways to train one, and it remains common practice. Some frontier teams have closed that route for their own models: what comes back is an answer, or a short summary, without the detailed reasoning trace behind it. That is a…
arXiv
2026-10-21
机器学习进展 (cs.LG)
While Chain-of-Thought (CoT) prompting improves the reasoning performance of large language models (LLMs), recent studies have shown that final answers do not always depend on the generated rationales. In particular, LLMs may rely on query-specific shortcuts rather than the information provided by t…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Large language models (LLMs) have shown great potential as general-purpose user simulators for interactive systems. Prior studies have demonstrated that training on high-quality data can substantially improve model reasoning and downstream task performance. However, the upper bound of user simulatio…
arXiv
2026-10-21
机器学习进展 (cs.LG)
While agentic reinforcement learning (RL) has emerged as a promising paradigm for training small language models (SLMs), existing frameworks are predominantly tool-centric and struggle with data-centric environments. Specifically, applying tool-centric frameworks to stateful tasks like SQL debugging…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Language models relying on verbalized Chain-of-Thought (CoT) are constrained to token-by-token generation in the language space, where many tokens serve textual coherence rather than core reasoning. This entangles genuine differences in logical trajectories with superficial variations in verbalizati…
arXiv
2026-10-21
机器学习进展 (cs.LG)
Quantization-aware training (QAT) aims to recover reasoning capabilities in low-bit large language models (LLMs). Existing recipes typically supervise the quantized student with the original full-precision model on fixed response sequences. However, the resulting low-bit quantized models still strug…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Through a tangle of real-world interactions, from synthetic training pipelines to data shared on the open web, frontier LLMs absorb what their predecessors and competitors create. How well does each model play its role in this ecosystem, as a creator and as a solver? We introduce Environment Duels,…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Quantitative research is an iterative loop: a researcher proposes a market hypothesis, implements it as an executable signal, tests it on historical data, and decides what to keep. Existing benchmarks score formulas, code, or individual trading decisions, leaving open whether large language models (…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Many NP-complete (NPC) problems have a polynomial-time (P) twin that differs in a single structural detail, such as 3- versus 2-colouring a graph. A model that reasons rather than recalls should notice which side of this boundary a problem falls on, yet existing benchmarks either fix the wording of…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Enterprise agents act on documents. Before an agent can follow a policy, answer a compliance question or file a ticket, some ingestion layer must turn a Word file or PDF into tokens - and that layer is where an agent deployed in the wild silently loses information. We show that one layer is lost by…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
LLM agents with similar success rates can follow different action sequences. An Agent Behavior Map represents LLM-harness combinations using centered action log-likelihoods on shared reference trajectories, without requiring new environment rollouts from the target agent. We derive error bounds rela…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Long-term memory for multimodal agents remains challenging, as it must preserve fine-grained visual evidence while tracking entity evolution across sessions. Existing abstraction-based memories compress images into text, often discarding visually discriminative details, whereas pixel-preserving memo…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Forming an accurate theory-of-mind (ToM) is key to successful multi-agent collaboration and requires reasoning under uncertainty to capture multiple possible partner states. This is hard to do in the text representation space native to large language model (LLM) agents, which more easily describe a…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Agent memory systems such as Mem0 and Zep decide what to keep before any question arrives; coding agents keep the raw history and let the model search it with its own tools. We ask whether designed memory still helps once the model is strong enough to search for itself. On one LoCoMo conversation (4…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Literature-searching language agents are becoming a standard component of scientific assistants for the life sciences, yet there is little controlled evidence on which parts of their harness matter. We run a controlled ablation on biomedical claim verification (SciFact) with the live PubMed API as t…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Most contemporary artificial intelligence (AI) systems are released with frozen parameters and improve through external updates, whereas an ideal system would keep improving from its own experience, a capacity we call recursive self-improvement (RSI). Yet studying RSI requires an environment that is…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Large language models (LLMs) are increasingly deployed as agents, yet it remains unclear how their foundational capabilities translate into agentic capabilities. End-to-end scores leave unresolved whether low benchmark capability or weak deployment drives a failure, limiting actionable diagnosis. We…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Retrieval Augmented Generation (RAG) grounds large language models in external knowledge. Modern pipelines combine embedding, chunking, retrieval, reranking, query augmentation, context compression, and agentic verification, yet these techniques are typically evaluated one at a time. It remains uncl…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Long-term agent memory is shifting from predefined workflows toward learnable policies for memory construction and retrieval. Existing learning-based methods either optimize construction alone or combine retrieval-side online reinforcement learning (RL) with offline construction supervision, limitin…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
LLM-generated records increasingly populate production information systems where they are treated as data assets subject to the same quality expectations as records from operational databases. A single such record can simultaneously satisfy and violate a quality rule: a grounded fact adjacent to a h…
arXiv
2026-10-21
大模型与智能体 (cs.CL)
Agent harnesses have become a central component of modern agentic systems, shaping how models use tools, manage context, and coordinate long-horizon interaction. Realizing the benefits of these harnesses depends on how effectively the underlying model can make use of them, motivating reinforcement l…
arXiv
2026-10-18
AI 与计算社会 (cs.SI)
The explore/exploit trade-off is the core problem of reinforcement learning (RL). We show that a minimal foraging swarm approaches its optimum - the ideal-free / Nash allocation of a multi-agent resource-sharing game (an equilibrium, not a return maximum) - with no learner and no reward signal. Agen…
arXiv
2026-10-16
AI 与计算社会 (cs.SI)
Newcomers to online communities must learn a group's norms to participate successfully. This process is known as socialization, and has been the subject of considerable study in the online communities literature. Foundational work on group socialization in social psychology, however, suggests that s…
arXiv
2026-10-16
AI 与计算社会 (cs.SI)
Generative search engines answer local and commercial questions with a written answer and a few cited sources. A growing practice of generative engine optimization (GEO) monitors which sources and brands these answers show, typically from a single place. In this paper, we present a controlled audit…
arXiv
2026-10-16
AI 与计算社会 (cs.SI)
Perceived gender inference from names is important for inclusive scientometrics research, yet existing methods perform poorly on non-Western scholars. We address this gap for Chinese names, which appear in two forms, Hanzi and Pinyin, its Romanized form (e.g., Zhang Wei, Li Na). In international bib…
arXiv
2026-10-16
AI 与计算社会 (cs.SI)
Graph-level anomaly detectors are trained only on normal graphs, yet benchmarks may use test labels for model selection. We examine whether UB-GOLD results reused by later papers support label-free detection. We audit UB-GOLD and 14 GLAD code bases, re-run 21 detectors on UB-GOLD's folds under selec…
arXiv
2026-10-16
AI 与计算社会 (cs.SI)
Decentralized Finance (DeFi) transactions involve complex interactions among accounts, smart contracts, and tokens. We investigate whether network topology can identify transaction methods that describe DeFi operations. We extract directed graphlet features to classify the 30 most frequent methods.…
arXiv
2026-10-16
AI 与计算社会 (cs.SI)
What the Web community believes about toxic conversations, such as that toxicity rises in longer and deeper threads, was measured mostly with Google's Perspective API, which is shutting down; large language models (LLMs) will make the next measurements. We find that LLM labelers mistake second-perso…
arXiv
2026-10-21
AI 安全与评测 (cs.CR)
Perimeter triage faces a latency/semantics trade-off: signature engines are fast and format-safe but blind to unseen attacks, while generative models are semantic but slow and produce unparseable output. In late 2026 a class of commercial System One decision models (TypeSafe Jev; Microsoft-Decision-…
arXiv
2026-10-21
AI 安全与评测 (cs.CR)
Malicious actors frequently obfuscate their code to hinder investigative reverse engineering efforts. Among the most formidable techniques is virtual machine (VM) obfuscation: programs are translated into bytecode for a custom VM, often with a randomized virtual instruction set and distributed along…
arXiv
2026-10-21
AI 安全与评测 (cs.CR)
Current LLM safety benchmarks predominantly rely on binary safe or unsafe labels, lacking explanatory depth and suffering from low consistency across automated judges. To address this, we introduce DiagSafe, a diagnostic evaluation framework that moves beyond binary classification by employing promp…
arXiv
2026-10-21
AI 安全与评测 (cs.CR)
Recent advancements in Vision-Language Models (VLMs) enable end-to-end visual grounding via language prompts and generative decoding, demonstrating largely improved zero-shot and open-vocabulary detection capabilities. As a result, VLMs are rapidly being deployed at the edge in safety-critical real-…
arXiv
2026-10-21
AI 安全与评测 (cs.CR)
The Regular Syndrome Decoding (RSD) problem is a structured variant of classical Syndrome Decoding and underlies a growing number of efficient cryptographic constructions, including zero knowledge protocols, pseudorandom correlation generators, and MPC in the Head signatures. Although RSD is general…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
Customized agent deployments define policies over tools, resources, and execution conditions. Providing these policies in the task agent's prompt does not reliably prevent violations. Training a guard requires examples of tool calls that comply with or violate those policies. We introduce AgentPolic…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
Adoption of artificial intelligence in K-12 schools has outpaced guidance for students, teachers, and administrators. Existing AI-literacy frameworks emphasize understanding and use; cybersecurity frameworks predate AI-specific threats, and neither defines AI security literacy as a competency domain…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
Modern LLM agents persist state across sessions via two channels: long-lived memory files silently prepended to the system prompt, and scheduled tasks waking the agent without user input. A one-time indirect prompt injection committed to either channel can re-fire across sessions long after the atta…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
Agentic AI extends large language models (LLMs) from generating responses to planning and executing complex workflows, substantially expanding their utility while exposing applications to unsafe actions, information leakage, and adversarial manipulation. Multi-agent systems further expand these capa…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
Local Retrieval-Augmented Generation (RAG) systems are vulnerable to indirect prompt injection and retrieval poisoning, where malicious instructions in retrieved content are mistaken for trusted input. We present nokast-secureRAG, a roughly 0.5B-parameter small language model that acts as a semantic…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
LLMs increasingly operate over mixed-sensitivity contexts, where they must reason over information ranging from public to confidential while disclosing only what the user is authorised to receive. Existing approaches enforce access control as a binary decision: either fully answering or refusing, wi…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
The UK, EU and US governments have set timelines for post-quantum cryptography migration of critical systems, with deadlines ranging from 2028 to 2035. This paper focuses on IoT-class devices and microcontrollers and the viability of using PQC to connect these devices in secure TLS 1.3 connections.…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
Large language models (LLMs) can improve honeypot realism by generating dynamic, context-aware responses, but powerful cloud models incur substantial cost and latency, while lightweight models may provide weaker deception. This forces practitioners to choose between operational efficiency and intera…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
When an LLM agent acts on behalf of a user, it is expected to act within the authorization scope the user grants. Existing agent evaluations largely measure task completion or harmful consequences and can therefore overlook actions that violate this delegation without producing conventional safety o…
arXiv
2026-10-20
AI 安全与评测 (cs.CR)
Machine learning benchmarks act as guideposts for model development and spur the creation of training datasets intended to improve benchmark scores. We introduce BenchDoor, a family of data poisoning attacks that exploits this incentive. Beyond preserving benign performance, BenchDoor requires bench…