Science & Mathematics

Knowing when they are reliable is essential for language models to support human-AI collaboration. Yet interaction feedback is not always complete or correct, and may mislead models' understanding of their own capabilities. We introduce LADDER to study how language models assess themselves under unreliable feedback. Our findings show that models can accurately report supplied performance information while still struggling to form trustworthy self-assessments from interaction experience. Apparently accurate assessments can obscure capability differences and do not necessarily predict future task performance. These findings suggest that trustworthy self-assessment should be evaluated against observed task performance and its predictive value. Our study provides empirical evidence of the limits of model self-knowledge and a basis for evaluating capability awareness in reliable human-AI collaboration.

Artificial synapses that integrate nonvolatile memory with broadband optoelectronic learning are essential for next-generation neuromorphic computing. However, their realization remains challenging because conventional floating-gate memories exhibit limited optical functionality and poor scalability. Here, we report a van der Waals ReS2/hBN/InSe floating-gate synaptic transistor that combines efficient charge storage with broadband optoelectronic learning in a single device architecture. The heterostructure incorporates InSe as a floating-gate layer for stable charge storage, hexagonal boron nitride (hBN) as a tunneling dielectric, and ReS2 as a photoresponsive channel, enabling carrier tunneling under electrical and optical stimulation. Through the combined operation of these layers, the device achieves a large memory window, an on/off current ratio exceeding 10^6, stable multilevel conductance states, and reliable retention over 10^4 s. Broadband visible-to-near-infrared optical excitation enables controllable synaptic weight modulation and the emulation of essential synaptic functions, including short- and long-term plasticity, paired-pulse facilitation, and pulse-amplitude- and pulse-duration-dependent plasticity. Optical potentiation and electrical depression further enable bidirectional modulation of synaptic weights. Beyond these synaptic functions, the device demonstrates higher-order learning through associative and discrimination conditioning inspired by the Pavlovian learning paradigm, as well as reconfigurable Boolean logic driven by electro-optical inputs. These results establish a multifunctional van der Waals floating-gate platform that integrates nonvolatile memory, broadband optoelectronic learning, and adaptive information processing within a single transistor, providing a promising strategy for compact, energy-efficient neuromorphic hardware.

Linear recommenders and topic models are similar in that they implicitly learn from co-occurrences of items. However, there has been little study of how these two model families are related. In this work, we show that, under a latent-factor model of the user-item interactions, the latent topics can be discovered from the weight matrix of a closed-form linear autoencoder. These topics provide a basis for interpreting the recommendations, and can be obtained directly from the trained recommender without additional training. Building on our analysis, we further propose novel item and user normalizations that prevent popular items and very active users from dominating the topics. Our experiments on six public datasets demonstrate that our method achieves significantly higher topic quality than every dedicated topic model and matches or exceeds the best linear autoencoder on all six datasets, while its recommendation performance is competitive with the best tuned recommender.

Humanoid robots are moving from laboratory platforms toward connected fleets that share workspaces, infrastructure, tasks, and knowledge with other robots and humans. As deployments grow, the operation and improvement of each humanoid depend on information exchanged across the network. However, existing research has largely studied humanoids as individual systems. This survey examines networked humanoid intelligence through four functions of information exchange: communication, coordination, collaboration, and social intelligence with human interaction. We first establish the foundations of networked humanoid systems, define the question addressed by each function, and connect them through a common factory scenario. We review the literature along six dimensions: interaction architecture, exchanged information, physical coupling, operating timescale, system scale, and embodiment. We separate results obtained with physical and simulated humanoids from those based on other robotic platforms or disembodied agents, because the latter do not fully capture humanoid-specific constraints such as balance, whole-body motion, contact, and physical interaction. We examine hardware, datasets, and benchmarks, and distinguish risks handled locally from those networking introduces, such as stale information or unsafe knowledge propagation. Finally, we identify open problems in control-aware networking, compact body-centric messages, physically grounded coordination, network- and embodiment-aware knowledge sharing, fleet-level social consistency, and accountable human oversight.

Reconstructing 1-m imagery from Sentinel-2 combines two difficulties. The input and the reference are observed by different satellites, which differ in spatial resolution, spectral response, radiometry, acquisition time, and local geometry, so no synthetic degradation operator links them. The required magnification of ten in each spatial dimension also far exceeds the factors of two to four at which remote sensing super-resolution is usually developed. We study this real-world $\times 10$ problem using YREB-SR, which pairs 2022 Sentinel-2 RGB composites with 2022 Jilin-1 RGB references resampled to a 1-m grid over the Yangtze River Economic Belt. The benchmark contains 11,559 geographically matched scene pairs, with 6,409, 2,135, and 3,015 assigned to training, validation, and testing, respectively. We develop CrossVAR, a visual autoregressive model that predicts the 1-m latent as a nine-scale hierarchy conditioned on the Sentinel-2 input and estimates the remaining quantization residual with a diffusion refiner run without injected noise. Complete 1600 $\times$ 1600 scenes are reconstructed in overlapping windows that start on the Sentinel-2 pixel grid, as every training patch does, and every window is decoded at the size on which the decoder was trained before the windows are blended with Gaussian weights. On all 3,015 test scenes, CrossVAR achieves 27.60 dB PSNR, 0.7725 SSIM, 0.1451 LPIPS, and 7.15 FID, the best value of each metric among the compared methods. These results establish a strong autoregressive baseline for large-scale cross-sensor reconstruction. To facilitate reproducibility, the source code and implementation are publicly available at https://github.com/SHANOAL/RSSR.

Visual captures provide detailed scene appearance, but converting them into interactive simulation environments requires object reconstruction, physical modeling, and correction of intermediate errors. We present SceneAgent, an agentic pipeline that coordinates these steps through simulator-based visual inspection and repair. The pipeline processes 3D Gaussian Splatting (3DGS), photogrammetry, LiDAR captures, AI-generated 3DGS scenes, and 4D videos of dynamic scenes, and combines object segmentation, physical parameter estimation, part decomposition, articulation, and deformable simulation. To support policy learning, SceneAgent generates diverse object variants ("digital sisters") and scripted demonstrations. On a cluttered bin scene from GraspClutter6D, SceneAgent raises F1 at 1 cm for every coding agent with scorable outputs both with and without its skills, and its reconstructions are better than SimFoundry: all agent configurations exceed SimFoundry's F1 score, with a best of 60.75% versus 6.9%. On three dynamic scenes from 4DCodeBench, SceneAgent raises the mean dynamic IoU of Gemini 4 Argon from 0.86% to 99.09%. We further demonstrate diverse simulated interactions and preliminary policy transfer to a real robot.

In high-stakes domains such as finance, critical information evolves through multi-party interaction. Recovering these changes and their sources is fundamental to reliable speech understanding. Existing speech benchmarks focus on transcription, local understanding, and turn-level state prediction, leaving joint decision evolution across speakers insufficiently evaluated. We introduce DecisionTraceBench, a speech state-tracking benchmark for Decision-Critical Information State Tracking (DCIST). It jointly evaluates decision event sequences, speaker attribution, scope, and temporal evidence, using trajectory-level F1 to measure complete event-chain recovery. Our dataset comprises Chinese financial dialogues involving two to four speakers, with 100.39 hours of synthetic audio and 8.61 hours of human-performed recordings. On a fixed cohort of human-performed recordings, five cascaded systems and two direct-audio models achieve DCIST scores of 5.73% to 21.35%. Ablations with speaker-labeled reference scripts and no temporal localization requirements yield a maximum score of only 38.72%. Audio perturbation experiments further show that overall transcription error rates do not adequately reflect complete trajectory recovery. These findings reveal limitations in jointly understanding information provenance and state evolution, providing a benchmark for evaluating decision-process understanding and decision traceability in high-stakes speech applications.

Learning and reasoning are both essential to intelligent agents, but have traditionally been treated as two separate stages: first learning a knowledge base from data (or taking one as given), and then verifying queries as if the knowledge base were complete. This separation has been shown to be inefficient when the downstream goal is to verify only a small set of queries. Moreover, it neglects the possibility that data and a knowledge base contain partial yet complementary information, requiring an agent to combine both sources to justify or refute a query. Motivated by the need to integrate learning and reasoning more tightly, we study the problem of query-driven learning for reasoning. More specifically, we extend prior work to the first-order setting, which is critical for practical applications as the number of ground atoms typically grows rapidly. We begin by asking when validating or refuting a universal query given partial data and knowledge base is even possible, and identify a necessary and sufficient condition called bounded concealment. We then propose a quasi-polynomial-time algorithm that solves the task at the lifted level. Finally, we establish soundness and completeness guarantees relative to space-bounded lifted proofs, along with quasipolynomial sample and runtime bounds, demonstrating that lifted query-driven reasoning can be efficient even when unrestricted knowledge-base learning is not.

Users are increasingly delegating tasks to AI agents, allowing agents to perform long-running, multi-step workflows on their behalf. However, agents may perform these tasks in ways that depart from users' expectations for how they should be carried out, leading to what we call agentic delegation failures. We leverage a large corpus of 163,487 Reddit posts, describing AI agent use, to develop MANDATE: a taxonomy of 23 failure categories organized into seven groups. A follow-up study with 73 AI agent users examined the relationship between prevalence and perceptions of severity and noticeability, finding that failures rated as more severe are considered harder to notice, and that prior experience with a failure is associated with greater noticeability. Finally, we assess the maturity of research on each failure and identify commonly encountered failures that have received little attention (e.g., agents running without making progress). Our findings inform opportunities to make delegation easier for users to oversee and better studied by the research community.

The Poisson approximation theorem states that when $m$ balls are tossed uniformly and randomly into $n$ bins, the distribution of the bin loads can sometimes be approximated by $n$ independent Poisson random variables of mean $m/n$. We show that when the balls are tossed via a random $\mathbb{F}_2$-linear map, the load distribution is still governed by this Poisson distribution, albeit in a (necessarily) weaker manner. This allows us to show that a random linear map behaves like a fully random map in the following ways: 1. $\textbf{Expected loads.}$ We show that the expected value of the $r$th largest bin is governed by the $r$th largest value among $n$ independent Poisson random variables of mean $m/n$. Specializing to the canonical setting of $m=n$ and $r=1$, our bounds show that the expected maximum load when $n$ balls are hashed into $n$ bins by a random $\mathbb{F}_2$-linear map is no more than the corresponding expectation for a fully random map $\textit{plus}$ $\frac{3}2 + o(1)$. Specializing to $m\gg n\log n$, we show that the expected max load is $(1+o(1))m/n$ in this regime. 2. $\textbf{Concentration.}$ When $m\gg n\log (1/ρ)$ balls are tossed into $n$ bins randomly, we expect with $0.99$ probability that $\textit{every}$ density-$ρ$ subset of bins will contain a $(1\pm o(1) )ρ$-fraction of the balls. We show that this remains the case if the $m$ balls are tossed using a random linear map, and our bound matches the $o(1)$-factor produced by the fully random case. This gives a stronger version of the Leftover Hash Lemma for linear maps. An implication of this is that all bins have load $m/n \pm O(\sqrt{(m/n)\log n})$ with probability $0.99$. We obtain these results by showing that the average of a log-potential on the bin loads is upper bounded by the cumulant generating function of a Poisson random variable of mean $m/n$.

Prediction markets have emerged as an important mechanism for aggregating information and trading beliefs about future events. The transparency of on-chain transactions has enabled investors to identify historically profitable wallets, commonly referred to as smart money, and replicate their trading activities through copy trading. However, despite the growing popularity of smart-money tracking, systematic research on these wallets' trading behavior, profitability persistence, and the economic feasibility of following their trades remains limited. To address this gap, we conduct a comprehensive empirical study of Polymarket, the most popular decentralized prediction market, by reconstructing wallet-level trading activities from on-chain transactions, position changes, market histories, and settlement outcomes. We first examine the distribution and persistence of historical profitability using statistical analysis and logistic regression, and further characterize heterogeneous trading strategies through multidimensional behavioral profiling and unsupervised clustering. Building on these findings, we develop an evidence-grounded multi-agent research framework that integrates wallet behavior, market conditions, and real-world information to generate interpretable and verifiable explanations of individual trades. Finally, we implement a real-time copy-trading system and evaluate its practical feasibility through order-book-aware paper trading and live on-chain execution. Our findings reveal substantial behavioral heterogeneity among historically profitable wallets and highlight the challenges of translating observed trading performance into replicable economic returns.

Low-Rank Adaptation (LoRA) has emerged as the dominant paradigm for parameter-efficient fine-tuning of large language models. Standard all-layer LoRA applies adaptation throughout the transformer, despite heterogeneous contributions from its components. In this work, we conduct a systematic mechanistic interpretability analysis to understand how LoRA affects model internals. We observe that LoRA's adaptation effect is highly non-uniform, with significant component-wise and layer-wise sparsity. Most critically, we find that important layers are highly task-dependent, with limited overlap across tasks. Motivated by these insights, we propose Adaptive-LoRA, a calibration-driven method that automatically identifies task-specific layers via frozen-base zero-ablation, enabling efficient layer selection without exhaustive search. Extensive experiments on Qwen2.5-1.5B across 11 diverse reasoning tasks demonstrate that Adaptive-LoRA outperforms full-layer LoRA while using only 30% of layers, achieving a $3.5\times$ reduction in adapter parameters. This work bridges mechanistic interpretability and practical efficiency, providing both fundamental insights and a principled approach for adaptive fine-tuning.

Effective collaboration requires language models to understand what others know and believe, and to act accordingly. Models that assume others share their information may omit necessary communication and undermine coordination. Many theory-of-mind (ToM) benchmarks assess mental-state inference through story questions, but correct answers do not establish that models can use these judgments in subsequent tasks. We introduce a multi-level evaluation built on existing, verifiable tasks. We extend household collaboration and calendar scheduling with asymmetric information while retaining their original constraints and validators. Independent interfaces assess world understanding, information access, beliefs, decisions, and execution. Controls supply personal evidence or correct beliefs to distinguish state recovery from state use. Across six model configurations, these abilities diverge. On matched scheduling cases, Claude recovers all agent beliefs correctly but produces subjectively optimal plans in only 67.9% of cases. In household collaboration, supplying beliefs raises Qwen-27B success from 55.0% to 75.0%, yet Qwen-4B answers every fixed supplied-belief source question correctly while completing only 20.0% of the corresponding continuation condition. These results show why ToM evaluation must connect perspective inference to independently verified decisions and task consequences.

Coding agents resend a growing context on every model call, and context policies are usually judged by the prompt tokens they save. Executed work also depends on call count and prefix reuse: a caching runtime prefills only the tokens after the longest cached prefix. A shorter prompt can therefore cost more if it rewrites cached history. Our quality-gated protocol splits executed prefill into logical prompt minus reused prefix and measures KV residency, time, and energy. On 450 localization cases derived from SWE-rebench V2, preloading ten retrieved blocks (Eager) and adding anchored history compaction (Compact) both pass a five-point non-inferiority gate. In a separate 900-cell study, Eager cuts model calls by 35.7% and the GPU-energy proxy by 33.2% against on-demand search (Grep); Compact saves nothing further. Enabling prefix caching leaves Eager's prompt volume similar (+8.3%; its 95% CI spans zero) but cuts its uncached prefill by 67.7%. Deleting history can turn a token saving into extra work. With mini-SWE-agent on 100 SWE-bench Verified instances, keeping only the last five observations (Mask-5) cuts Eager's prompt volume by 33.3% but loses 135.9k reusable tokens per task. Uncached prefill rises by 608.5%, the energy proxy by 125.6%, and the policy misses the quality gate. Policies thus save physical work by shortening the trajectory and can multiply it by editing reusable history.

Generative retrieval (GR) typically relies on item-level beam search to retrieve candidates, leaving cross-item dependencies unmodeled and limiting direct optimization of list-level objectives. A natural alternative is to generate the entire list within a single autoregressive trajectory. However, greedy decoding with directly trained list-wise models substantially underperforms beam search. We identify two contributing factors: the broader search space explored by beam search and the positional bias of scarce positive signals, which deprives later positions of adequate supervision. To address these challenges, we propose $\textbf{SLIDE}$, a search-to-list item distillation framework that decouples offline search from online list generation. A teacher model retrieves high-quality candidates through offline item-level beam search, providing additional conversion-relevant supervision beyond sparse behavioral labels. Their Semantic IDs (SIDs) are then concatenated in beam-score order to form list-wise distillation targets. This search-to-list conversion transforms candidates retrieved along separate search paths into sequential item-generation targets within a single trajectory. Trained on these targets, the student generates the Top-$K$ list in a single greedy pass, also decoupling teacher scaling from online deployment. Experiments on FORGE and Taobao Search show that SLIDE substantially improves retrieval quality under single greedy decoding and, with stronger offline teachers, surpasses the beam-search reference. Supplementary RL experiments support the effectiveness of list-level generation.