ALT Multi-agent systems (MAS) require both effective coordination and specialized task execution. Existing prompt optimization approaches typically couple these functions within agent-specific prompts. We introduce Coordinator-Executor Prompt Optimization (CEPO), a general framework for improving prompt optimization in MAS. CEPO separates system-level coordination from agent-level execution through a shared coordinator prompt and agent-specific executor prompts, and guides their updates according to their respective roles. This structure allows shared coordination rules to be refined consistently across agents while keeping execution instructions separately adjustable. CEPO provides the prompt structure and optimization guidance, while existing prompt optimizers perform the underlying search. Extensive experiments spanning 11 benchmarks across four domains demonstrate that applying CEPO to four existing prompt optimizers consistently improves their performance. Ablations and qualitative an

ALT When a Mixture-of-Experts (MoE) model does not fit in GPU memory, batch-1 decoding is dominated by copying expert weights from host memory. Cache-aware routing reduces these copies by biasing the router toward resident experts, but combined with asynchronous prefetching it makes the model's outputs depend on transfer timing, so a harness that does not simulate that timing no longer evaluates the deployed model. We propose prefetch-steered routing (PSR): routing conditions on a deterministic logical cache, into which at most n experts predicted by the next layer's router are inserted per layer and token; physical copies affect only latency. On OLMoE-1B-7B with 50% and 25% of experts resident on an RTX 4090, PSR decodes 2.2-2.3x faster than exact routing with LRU and 25-30% faster than the Cache-Prior operating point, with mean benchmark differences of +0.1 and +0.5 points (95% lower bounds -0.8 and -0.3). In greedy generation its lead over Cache-Prior depends on the workload and reverse

ALT AI agents must decide when another search is worth its cost. Answer-success probabilities capture question difficulty, while this decision requires estimating the additional benefit of search. We introduce SearchHorizonBench, a benchmark for learning and evaluating search value from execution experience. Paired continuations and repeated executions separate expected search gain from generation noise and connect prediction quality to decision utility. The resource contains 45,632 executions across 3,168 questions. On 1,632 held-out questions, historical calibration reduces Brier error by 16.2%, with the improvement concentrated in success-level information. Search gains have a standard deviation of 21.4 percentage points, and tested historical estimators explain less than 2% of this variance in point estimates. Matched-configuration learning curves probe this distinction across training scales. At the reference token price, free selection observations yield a 2.49-point utility opportun

ALT Learning high-level causal variables from low-level observations is a central problem for artificial intelligence. With known node semantics but an unknown graph, visual readout errors can misalign model settings with environment intervention values. We address this mismatch by pairing image readouts with verified values attained through fixed round-robin interventions. With encoder and decoder frozen, a dimension-wise invertible calibration map puts visual responses and target settings in shared numerical units. After freezing this map, the agent selects interventions to resolve graph directions and fits local mechanisms to predict consequences; inverse calibration returns counterfactual states to the decoder's original coordinates. Across Domino, Pick-and-Place, Balance, Gears, CAUSAL3D, and CausalWorld, we evaluate graph discovery, intervention consequences, and counterfactual images. Controlled domino experiments show that interventions resolve orientation ambiguity left by equal a

ALT AI systems are increasingly present in children's everyday environments: education, entertainment, social networks, and toys. Their safety is typically assessed at the model level, through benchmarks, red-teaming, and content filtering. However, most AI harm taxonomies lack child-specific evidence and under-specify developmental harm and the stakeholder asymmetries unique to minors. Building on prior taxonomies, we construct a child-centered harm taxonomy mapped to sociotechnical failure modes. We curate and code 1331 publicly reported child-related AI incidents, annotating internal and external harm types and the failure mode implicated in each incident. Our analysis reveals which harms are most prevalent in public reporting, how individual-level and population-level harms relate, and which failures are most often coded alongside harm. We find that harm concentrates in representational, dignity-related, and psychological categories, and co-occurs most often with deployment-context, go

ALT In this work, we propose a movable-element (ME) multi-functional reconfigurable intelligent surface (MF-RIS) architecture, termed MM-RIS, which integrates element mobility with adaptive signal reflection, transmission, amplification, and energy harvesting (EH) to enhance communication flexibility and energy self-sustainability. A hybrid rate-splitting multiple access (RSMA) and non-orthogonal multiple access (NOMA) transmission scheme is further designed to improve spectral efficiency. We formulate an energy efficiency (EE) maximization problem by jointly optimizing transmit beamforming, power and rate allocation, MM-RIS configurations, panel deployment, and ME positions. To address the resulting non-convex mixed continuous-discrete optimization problem, we propose constrained hybrid deep reinforcement learning with federated transformer and parametrized sharing (CRAFTS). CRAFTS integrates multi-agent proximal policy optimization (MAPPO) and multi-agent deep Q-networks (MADQN) through

ALT Group fairness in retrieval-augmented generation (RAG) systems is typically evaluated using metrics aggregated across queries. However, such metrics can conceal topic-level disparities: a system may favor women on questions about human rights and men on questions about actors and filmmakers, appearing nearly neutral overall even though neither topic is fair. We therefore study topic-level fairness, which requires group fairness to hold within each topic rather than only in aggregate. To reduce the topic-level disparity, we adapt the reranker separately for each topic with a Bradley-Terry objective under direct supervision from the generated output: preference pairs built from the generator's own output carry the direction that topic requires. With preference pairs built from RoMQA alone, the adaptation reduces the mean absolute per-topic disparity by 51% on RoMQA and by 40% on TREC Fair Ranking while retaining accuracy. No baseline reaches the reduction, even when modified to treat eac

ALT Hierarchical task knowledge enables efficient problem solving by allowing previously acquired skills to be reused across tasks and problem instances. In this paper, we investigate curriculum-guided acquisition of hierarchical task knowledge in an HTN learning framework. We compare three approaches: online learning during problem solving, human-designed curricula, and an automated curriculum generation procedure that derives learning experiences from verified LLM-generated solution traces. The proposed procedure solves small training problems with an LLM, extracts task accomplishment information, and transforms it into ordered sequences of state-task learning problems for HTN method acquisition. We evaluate the three approaches across multiple planning domains. The results demonstrate that automatically generated curricula recover much of the benefit of expert-designed curricula while substantially reducing the token cost of online hierarchical knowledge acquisition.

ALT Goal reasoning (GR) is the process by which intelligent agents continually reason about the goals they are pursuing, which may lead to goal change, for instance when opportunities arise, or when events cause discrepancies among expected and actual outcomes of actions planned to achieve goals. In this paper we investigate whether Hierarchical Task Network (HTN) planning can provide a uniform implementation of the Goal Lifecycle, a flexible process model for goal reasoning in dynamic environments. The Goal Lifecycle progressively refines goals by adding constraints to them through a series of strategies (e.g., formulate, select, expand) and uses other strategies (e.g., defer, re-expand, repair) to resolve the current goal when an event impacts it. We present a novel GR agent architecture that employs HTN planning in this Goal Lifecycle. In addition to using HTN planning for the expand and re-expand strategies, as is frequently done in GR and planning research, we also use it in the formu

ALT Effective tutoring requires determining not only what information to provide, but also which form of assistance to deliver. A tutor may address the same student error using various pedagogical strategies, adapting its approach as the interaction progresses. We introduce Comparative Pedagogical Supervision (CPS), a training framework to learn to select the optimal strategy and feedback for student learning. We evaluate CPS on math and science reasoning datasets. After five tutoring turns, CPS outperforms the strongest fixed-strategy baselines. The learned tutor transfers across datasets, and its strategy selection can also inform a separate API-based tutor.

ALT AI agents are increasingly proposed as autonomous experimenters, yet the benchmarks that test them mostly offer cheap experiments that answer at once. Experiments in physical facilities are different: they take months, occupy scarce equipment and cannot be undone. We present SlowLab, a benchmark in which an agent runs a one-year campaign in four simulated greenhouse compartments on a site whose crop, facility and prices are hidden. The agent decides when to plant which climate policy, whether to stop a crop and which measurements to read while crops grow, and at the end recommends one policy, which a private evaluator scores in weather the agent never saw. In a preregistered study on 48 unseen sites, three low-cost language-model agents ended near a standard grower setting and clearly below a Bayesian-optimisation baseline that starts from that setting. They ran the facility as two synchronous batches, so measurements taken during the season never changed what they planted, and these m

ALT Large language model (LLM)-based personal agents are increasingly deployed to handle everyday tasks on users' behalf. These agents have access to private user information, while their tasks involve communicating with third parties; this raises a critical privacy concern that agents may inappropriately reveal sensitive user information to these recipients. Existing benchmarks primarily assess direct privacy leakage, where an agent explicitly states a pre-specified secret or attribute. However, they do not capture inferential leakage, where the model leaves indirect cues in the reply that implicitly leaks sensitive attributes. We introduce textscInferPriv, a benchmark that measures whether an honest-but-curious recipient can infer a user's sensitive attributes from an LLM's reply on routine task. Evaluating five frontier models, we find that inferential leakage substantially exceeds direct leakage across models and prompts. Under a privacy-aware system prompt, GPT-6.1-Sol, Sonnet-5.5, an

ALT Text injection trains speech recognition models on unpaired text alongside paired audio. We feed the speech encoder synthetic input derived from text so the whole model learns from it. Previously, we synthesized this input with a frozen text-to-speech (TTS) model. Here, a simple lookup table replaces the TTS model: a mean log-mel vector (or one-hot encoding) per phoneme, repeated for a random duration drawn from alignment statistics, with silence between words. Training is 1.8 times faster than with the neural TTS, achieving comparable word error rates. With an external language model, our system reaches 1.36 / 2.90 / 1.52 / 3.24 % WER on LibriSpeech dev,test-clean,other, to our knowledge state-of-the-art using only LibriSpeech paired data, text, and lexicon. We analyze components of synthetic input (durations, acoustic features), compare different HMM aligners, evaluate scaling on the 25 000 h Loquacious dataset, and show that full-encoder injection yields strong relative improvements

ALT Indoor context deciphered by ambient sensing underpins many applications, from assisted living to smart buildings, yet remains weakly connected to the language-aligned paradigm of modern multimodal systems. Through billions of IoT devices, ambient WiFi Channel State Information (CSI) captures the indoor context of human activity and environmental dynamics passively via ubiquitous, always-on infrastructure. However, unlike vision or audio, CSI lacks naturally paired text like captions or transcripts, limiting its integration with language-driven AI systems. We present CSI-TextBench, a large-scale dataset and benchmark for aligning CSI data with natural language, unifying 3 sensing benchmarks into 285,790 samples across 7 tasks and pairing them with 2,291 hierarchically structured, label-derived descriptions via many-to-many assignment to construct 2.0 million CSI-text training tuples. Benchmarking 5 alignment methods from 4 paradigm families, we show that contrastive CSI-text alignment

ALT We release a battery of 28 propensities for AI research, with expert-written rubrics, synthetic datasets, and annotated demand intervals, built with the method of Romero-Alvarado et al. (2026). We selected a wide collection of generic propensities that may influence the behavior of AI systems and humans. They are inspired by psychometric, psychological, and behavioral research, so that experimental results can be understood in the context of familiar constructs previously used also for human behavior research. The rubrics are useful for annotating existing benchmarks whose items may be relevant or not for one or more propensities. For easier profiling, however, the synthetic datasets are defined in such a way that they should be sensitive to one propensity only, allowing for a quick prototyping of propensity profiles using the battery of datasets. We also outline directions for future propensity research.

ALT We describe HobbyLM, a family of small sparse Mixture-of-Experts (MoE) models trained from scratch with a few hundred to a few thousand dollars of rented GPU time. The core is a decoder-only transformer with fine-grained experts, sigmoid routing with auxiliary-loss-free load balancing, grouped-query attention with per-head QK-normalization, and aMuon/AdamW optimizer split. Each of these choices was selected with a single-variable ablation suite at 130M parameters. On this architecture we train a 500M-parameter model (169M active) on 40B FineWeb tokens and post-train it into a chat model, a masked-diffusion language model, a function-calling and computer-use agent, and a single "omni" checkpoint that reads text, images, video, sound and speech through frozen encoders and small projectors. We run these models on a laptop CPU with a from-scratch Rust inference engine. Two of the authors then adopted the architecture without architectural changes to build HobbyLM-1B (1.04B total, 305M acti

ALT Statement autoformalization translates mathematics from natural language into a proof assistant such as Lean 4, and models now do it at scale. As provers get stronger, the burden of correctness shifts from the proof to the statement, that is, to whether the formal statement faithfully translates its source. Given a trusted reference formalization, the natural test is logical equivalence, and it proves too much. P ↔ Q holds trivially for any two statements P and Q that the library proves, because a prover can prove each separately and pass through P ↔ True ↔ Q. The stronger the prover, the more unrelated pairs pass. Earlier methods restrict the prover and check that each proof names the other statement. We introduce SemantEq, a more principled relation that instead requires each statement to turn into the other only through a chain of allowed steps, such as logical rearrangement and rewriting with the library's own equations. Our programs judge a proof by the proof term Lean checks, not

ALT While multimodal contrastive models like CLIP and CLAP successfully learn shared representations, their embeddings typically exhibit a persistent modality gap. We propose addressing this separation through a purely geometric approach: changing the embedding domain from the unit sphere, Sᵈ⁻¹, to the probability simplex, Δᵈ⁻¹. Although both manifolds share the same intrinsic dimension, the simplex acts as a structural prior; its nonnegative, unit-mass constraints serve as a cross-modal calibration that forces both modalities to ground themselves in the same coordinate system. By attaching lightweight softmax adapters and using Total Variation similarity, we reduce the centroid ℓ₂ gap by 97–99% across five multimodal retrieval benchmarks. This confirms that simplicial geometry induces genuine semantic alignment, matching or exceeding the retrieval performance of standard unconstrained models.

ALT Reinforcement learning (RL) post-training improves the reasoning capabilities of large language models, but its cost is dominated by autoregressive rollout generation, which can exceed 70% of training time. Existing speculative-decoding methods are poorly suited to this setting: model-based drafters consume scarce GPU memory and may become misaligned as the policy evolves, while existing model-free methods do not explicitly exploit repetition across training epochs. We observe that recent rollouts of the same prompt provide a high-quality, self-updating draft source. Based on this insight, we propose NgramRL, a per-prompt n-gram drafter constructed from bounded recent rollout history. NgramRL requires no auxiliary model, drafter training, or persistent GPU-resident drafter state, while target-model verification preserves the rollout distribution. To hide CPU cache lookup and scheduling, we further introduce an adaptive fan-out prefetcher that predicts likely verification outcomes (j, b

ALT In LLM-agent opinion simulations, each agent reports the probability that a claim is true after seeing its own previous probability and the probabilities reported by its neighbors, and that number becomes its state in the next round. If the number describes the agent's position rather than the sentence it was shown, then rewriting the claim as its complement and replacing every probability p with 1 - p should return the complementary answer, so the two reports should sum to one. We test this with three small open models: Qwen3-4B, Llama-3.2-3B and Gemma-3-4B. On 2,304 agent states taken from a running network simulation, Qwen's paired reports sum to 1.18 on average, and on 48 real fact-checking claims from AVeriTeC they again exceed one. A controlled benchmark of 24 worlds, each with exactly two possible outcomes, shows the same failure in all three models. The discrepancy is two to ten times larger than the change produced by a same-meaning paraphrase, and it persists under repeated s