Papers

Verifier-based selection improves LLM performance by generating multiple candidate solutions and using a verifier to select the most promising one. However, existing methods typically treat verification only as a ranking step and discard its feedback once a fixed candidate pool has been evaluated. In this paper, we ask whether verification can also improve the candidate set itself. To this end, we introduce LLM-as-an-Improver and propose Verify–Repair–Reselect (VRR), which uses verification feedback to generate and reselect improved candidates. VRR retains the initial winner while conditionally generating three complementary alternatives: repaired versions of the winner and runner-up, and a solution based on a new approach. It filters invalid and duplicate candidates using only inference-time information and then reselects the final answer under the original evaluation criteria. Across diverse models and code-generation and reasoning benchmarks, VRR improves over fixed-pool verifier-based selection in many settings and can recover correct solutions even when all candidates in the initial pool are incorrect. These results highlight a broader role for LLMs as improvers: verification feedback can not only select among existing solutions but also construct stronger candidates beyond the initial pool.

High-quality pre-training data is a critical bottleneck for educational and STEM-specific language models targeting edge AI and on-device deployment where token budgets are tightly constrained. While major organizations train ever-larger models on private corpora, the open ecosystem lacks STEM-focused synthetic datasets that deliver high per-token learning value efficiently for small models. To address this gap, we introduce QVAC Genesis III, a 191.43B-token, STEM-focused multi-domain synthetic corpus covering 19 domains across several difficulty levels and different educational styles. QVAC Genesis III is built via a dual generation strategy that performs targeted teacher distillation using a weak edge-scale student model as signal: the student's failures are converted into corrective explanations, while its successes are expanded into contrastive option-level reasoning over all answer choices. We further introduce an LLM-as-a-parser evaluation protocol that extracts final answers from free-form outputs and tracks both accuracy and answer validity. To validate the effectiveness of our QVAC Genesis III data, we conduct controlled from-scratch ablations with 1.7B-parameter models, showing that models trained with QVAC Genesis III consistently outperform both models trained with the open-source synthetic corpus Cosmopedia-v2 and the publicly released Cosmo-1B model across ARC, GPQA Diamond, and MMLU STEM benchmarks, achieving up to +28.57% on ARC-E and +21.35% on ARC-C, while reaching a Valid Answer Rate of up to 99.45%.

Robots can recall prior failures without knowing whether recalled evidence remains valid, conflicts with current observations, or is sufficient to guide a decision. We present CoreSense, a robot-system integration architecture that combines traceable episodic evidence with a conflict-aware belief gate and bounded, auditable recommendations. The gate checks scope, provenance, time, contradiction, and support before it permits PROCEED, requests re-observation, abstains, or escalates. Evaluation follows three complementary layers without commanding a physical robot: offline public real-robot data, a frozen signal-level simulation, and a live cloud deployment path. On CableTrace-120 and BotFails-200, belief gating reduces protocol-defined unsafe proceeds from 20% and 40% to 0%. A disjointly calibrated raw-video policy also reaches 0% unsafe proceed, but overblocks every nominal episode. On public data, a ViFailback-BotFails visual detector reaches 0.778 AUROC yet remains all-blocking, whereas cycle-disjoint UR3 telemetry for protective stops yields 0% unsafe proceed, 36.1% overblocking, and 61.9% coverage; grip-loss transfer remains a negative result. Controlled physical corroboration yields 3.3%, 0%, and 42.0%, while conflict-aware fusion yields 4.7%, 0%, and 42.8%. Finally, 20/20 cloud recalls validate a CockroachDB Cloud-Amazon Bedrock deployment path. The evidence supports an auditable integration pattern, not autonomous recovery or certified safety.

Constructive classifiers add structure while they train: a level to a tree, a unit to a hidden layer, a split at a leaf. This paper asks what each of four such growth decisions buys, measured under one protocol on 24 datasets, and gives an exact diagnosis and fix for the one that buys nothing. The diagnosis concerns the natural way to deepen a soft decision tree: turn every leaf into a gate whose two children inherit the parent's class distribution, so the function is unchanged. I prove that this leaves the gradient of every new gate identically zero and with the gate at 1/2, gives the two children identical gradients, so under this construction the added level can never learn. That predicts where it costs: nothing on two-class problems, where two leaves already suffice and a great deal where more classes need more leaves. Measured, the cost is -0.2 points over 13 binary datasets and 40.2 points over 7 multi-class ones, and it tracks the number of classes (Spearman 0.64), not the number of features (0.03). The fix is any perturbation of the children. Neither its size nor its direction matters: a residual-directed initialisation changes accuracy by +0.20 points against noise. The other three decisions each buy one thing. Fitting a new hidden unit to the residual before installing it buys a smaller network but not a more accurate one. Splitting one leaf at a time buys sparsity; it lost accuracy until I found that the split started its two children identical and untrained, the same defect in another place; with the children inheriting the parent and the symmetry broken, per-leaf growth comes within 2.3 points of a complete tree using 23% of its splits. Requiring statistical significance before a node gets a more expressive split buys nothing. Every number comes from the measurement scripts.

While neural theorem provers have achieved impressive milestones in formal mathematics, they largely operate on the assumption that faithful Lean 4 formal statements are already provided. Translating informal natural language into a formal language is a critical data bottleneck plagued by an "illusion of rigor": standard type-checkers accept statements that compile but drop hypotheses, introduce vacuous truths, or subtly alter mathematical bounds. To resolve this, we introduce Sage (Semantic Agent-Guided Formalization Engine), an agentic framework that replaces monolithic translation with a four-stage decomposed generation pipeline coupled with a dual-signal semantic correction loop. By pairing Lean 4 compiler diagnostics with multi-dimensional semantic feedback, our correction loop enforces mathematical fidelity alongside syntactic validity. By explicitly accounting for the gap between open-ended queries and declarative formal targets, our pipeline prevents models from achieving high formalization rates by guessing unverified answers (exhibiting a 70.9% answer leakage rate in monolithic baselines). Consequently, Sage suppresses leakage to 2.7% while achieving 73.3% pass@4 joint compilation and semantic fidelity on the Omni-MATH without proofs (compared to 42.0% for a fine-tuned Goedel-Formalizer-V2 baseline). Finally, on IMO-Unformalized, a novel frontier of 175 unformalized International Mathematical Olympiad problems, Sage demonstrates effective zero-shot generalization with 87.4% pass@4 verified fidelity compared to just 19.4% for the baseline, winning over 79% of blind pairwise evaluations.

Diffusion language models (DLMs) enable flexible, parallel generation, but their iterative denoising remains computationally expensive, motivating increasingly aggressive compression. Existing compression objectives largely measure how well compressed computation approximates the original locally, but local error does not reveal which removed computations actually matter to the downstream denoising trajectory. We introduce Same-State Causal Restoration (SSR), which restores the original FFN on the exact current input reached by the compressed model and measures how the resulting trajectory changes. To our knowledge, this is the first direct measurement of the same-current-input closed-loop effect of removed FFN computation in DLM compression. Across LLaDA-8B-Instruct and Dream-v0-Instruct-7B, compressed-side state ranks this downstream effect substantially better than local NMSE at fixed denoising phase, while controlled interventions show that correction structure matters beyond magnitude. Using task-label-free calibration, SSR freezes a single restoration window for held-out inference. Under aggressive LLaDA compression, restoring only four transitions recovers 89.9% of the lost accuracy while retaining an estimated 36.8% whole-model MAC saving and outperforming an equal-budget local-error baseline. Dream further shows that restoring dense behavior and repairing the final task are distinct outcomes.

This work presents an end-to-end demonstration of autonomous machine learning research conducted by an agent skill on consumer hardware. The demonstration evaluates an FFT-based Kernel Ridge Regression (KRR) solver for regular spatial grids using 2005 monthly NOAA Kaplan SST v2 anomaly fields on a $36 \times 72$ grid. This was autonomously executed by DeepSeek V4 Flash, orchestrated by our agent skill suite within DeepSeek Harness (DSH). Experiments were executed on CPU-only hardware (Apple M2 Pro; 78.7 s solver time, 1.57 GB peak RSS). Long-horizon state was decoupled into a file-based epic- and issue-tracking substrate. Across 74 sub-agent sessions, the agent demonstrated closed-loop scientific resilience: routing two failed hypothesis review gates back to literature retrieval, patching bootstrap indexing bugs, and executing with only four discrete human steering events. Finally, we reflect on autonomous research governance, arguing that scientific credibility requires inspectable state, falsifiable review gates, and transparent reporting of negative results, urging the machine learning community to favour agent-accessible structured formats over static PDF manuscripts.

Joint-embedding predictive architectures (JEPA) have emerged as a promising self-supervised pretraining paradigm for time series, learning representations by predicting target embeddings in latent space rather than reconstructing raw signals. Yet evidence on their benefits remains mixed, and most studies test only a single backbone or a narrow set of architectures, leaving unclear whether JEPA pretraining is a reliable improvement or one that depends heavily on the downstream model. We address this gap through a large scale evaluation of one JEPA instantiation across nine backbones and eleven benchmarks spanning temporal and spatio-temporal forecasting, the most extensive cross architecture assessment of JEPA for time series to date. We find that the benefit of this instantiation varies sharply across backbones, producing consistent gains for some architectures and consistent degradation for others, even on the same dataset. This pattern holds across both task families, indicating the variability is a general property of this instantiation rather than a dataset specific artifact worth accounting for when choosing a backbone in practice.

Coccidiosis caused by Eimeria parasites is a major economic burden in poultry production, and effective control depends on accurate species-level diagnosis. This study evaluates whether a general-purpose multimodal large language model can support such diagnosis. Google Gemini was assessed on 4,225 mi- croscopy images covering the seven fowl-infecting Eimeria species under two prompting conditions, one without candidate labels and one with a predefined class list, and was further tested for pathology-report generation, while BiomedParse was examined for parasite segmentation. Without candidate labels, the model produced broad and taxonomically inconsistent outputs. With candidate labels, overall accuracy reached only 14.9%, with a strong bias toward E. tenella at 74% and no correct classifications for E. acervulina, E. mitis and E. praecox. Generated treatment reports were coherent but unverified, and segmentation was only partial. Current multimodal models are therefore not yet reliable for standalone Eimeria diagnosis without domain-specific fine- tuning and expert validation.

Face recognition is widely deployed in surveillance, access control, and forensic workflows, yet accuracy degrades sharply once the face is occluded by accessories, foreground objects, or the frame edge. Because most faces in the wild are partial, robust partial face recognition (PFR) remains open. This paper compares three pretrained convolutional backbones, ResNet-50, VGG-16, and FaceNet, fine-tuned for PFR by transfer learning under identical preprocessing, splitting, and optimization protocols on the Specs-on-Faces (SoF) dataset. All three arms use a common 160x160 input and a frozen backbone with a trainable head under a fixed epoch budget and no per-backbone hyperparameter search. The FaceNet configuration, denoted PFN (Partial FaceNet), substantially outperforms the other two, reaching 97.4% test accuracy with macro-averaged 87.04% precision, 84.61% recall, and 84.17% F1 over the 112 identity classes, the highest accuracy and recall reported on SoF.

We introduce the Active Causal Discovery Benchmark (ACDB), an SCM-grounded environment for evaluating whether LLM agents recover causal graph structure from observations and budget-constrained hard interventions. ACDB pairs a linear-Gaussian world generator with a fixed observe-intervene-submit API and a three-layer scoring contract that separates skeleton recovery, DAG recovery, and intervention efficiency. On the current six-level ladder, PC with a greedy active orientation heuristic is the strongest non-oracle method (directed F1 42.7%, SHD 4.79), ahead of Claude Sonnet 4.6 raw active (31.7%, 7.25) and GPT-5.4 raw active (22.9%, 9.27). The most informative diagnostic is the precision-recall decomposition: PC under-commits with high precision, LLMs over-commit with lower precision, and statistical-tool access often increases abstention rather than useful intervention. A structure-blind random DAG baseline reaches 23.6% directed F1 on this dense v0 ladder; a density probe lowers this floor to 16.9%, motivating the v1 calibration pass. The current results should therefore be read as a benchmark audit and calibration report, not as evidence that current LLMs solve active causal discovery.

Conversational recommendation agents are a new paradigm for content discovery, enabling users to express complex intents through natural language (e.g., "recommend Italian indie artists I haven't heard before"). A central challenge in building such agents is optimizing agent planning – deciding how to select, sequence, and invoke tools – particularly in cold-start settings where real user interactions are not yet available. We introduce a pipeline for multi-turn synthetic data generation and a self-improvement loop to address this challenge. The synthetic data pipeline transforms single-turn prompts into realistic multi-turn conversations, enabling systematic evaluation before launch. The self-improvement loop combines variance-based contrastive optimization with iterative refinement through a coding agent, automatically identifying and fixing planning and tool-use errors. Our approach improves quality by +8% on top of a highly optimized manual prompt. The system has been productionized and significantly accelerated iteration cycles for the launch of a conversational recommendation agent at Spotify. Online A/B tests demonstrate its effectiveness, with +14% user listening, +5% increase in weekly active users, and a 5% reduction in skip rate compared to a prior experience supporting only session refinement. This work provides a practical framework for accelerating the development of conversational recommendation agents in industry.

TW3Cast is a time-series forecasting system that reaches position 3 of 130 entries on the GIFT-Eval benchmark by mean MASE rank, as of 2026-09-14. The two entries above it belong to the leaderboard's agentic category, multi-step systems that use agents or language models to reason about, generate or select forecasts. TW3Cast runs no agent and no language model. Its selection is a table computed once on the training split and then frozen, and its experts are public foundation models lightly fine-tuned on those training splits. For each of the 97 dataset, frequency and horizon configurations, the table serves one of four modes: a specialist, which is a LoRA or full fine-tune of Chronos-2, TiRex or Toto whose training data was cleaned and enriched by explicit rules; a quantile blend that contains a specialist; a blend of base models; or a selection tournament played on a backtest carved from the training split. Every decision in the table was taken on that backtest. A specialist is admitted the moment it beats the tournament there, so a candidate costs a few megabytes and minutes of GPU time, and a failed candidate changes nothing. Three guarded mechanisms protect the selection from its own biases: a dual accuracy and calibration criterion, an asymmetric margin against candidates that saw the series during training, and conservative per-window gates. The selection rules themselves were chosen inside a temporal meta-backtest. The best base model served alone reaches a mean MASE rank of 33.8, the tournament served on every configuration reaches 38.0, and the full router reaches 19.4. The routing table, the expert index, the pinned base-model revisions, the submitted score file and the dated snapshot of the public scores are released, and every leaderboard number in this paper regenerates from them by one script.

This research addresses the challenge of limited material recycling in wood recycling processes by leveraging artificial intelligence (AI)-based dimensionality reduction. Our study explores the application of convolutional neural networks (CNNs) in multi-channel hyperspectral imaging (HSI), extending beyond RGB channels to over 200 spectral channels. Dimensionality reduction within this context involves streamlining the feature space for AI system training and inference. Focusing on explainable AI (XAI) methods, this paper contributes to a broader research initiative, presenting a solution framework that enhances the sustainability and efficiency of wood recycling processes.

Generalizable robot manipulation requires predicting how a scene will evolve, identifying where interactions are feasible, and determining how to act. Action-labeled robot videos directly supervise control but are costly and limited in diversity, whereas egocentric human videos capture diverse interactions but lack robot actions and differ in embodiment and appearance. We introduce AffordanceWAM, an affordance-aware generative World Action Model that represents object-centric spatiotemporal affordance through Scalar Affordance and Affordance Heatmap, within the generated future World. This representation grounds visual prediction in task-relevantant objects and interaction regions for action generation, and provides shared interaction targets across human and robot videos. Built on a pretrained video diffusion Transformer, AffordanceWAM uses separately parameterized World and Action Experts, coupled through Masked Joint Self-Attention, to jointly predict future RGB observations, Scalar Affordance fields, Affordance Heatmaps, and continuous robot actions under a unified flow-matching objective. Human videos supervise all three future-World streams, whereas robot trajectories additionally provide action supervision, enabling transfer without human action labels or retargeting. Experiments on RoboCasa, CALVIN ABC$\rightarrow$D, and real-world manipulation demonstrate consistent gains over RGB-only and robot-data-only baselines. Under fixed robot supervision, RoboCasa performance improves monotonically as affordance-annotated human video scales. These results support affordance as an effective interface for both vision-language-action learning and human-to-robot transfer.

Black-Box Optimization (BBO) is often applied in several engineering fields and can utilize an advancement of numerical measure- ments and simulation technologies. It deals with the optimization func- tions, where an analytical description is unavailable. It relies on meth- ods that require only an input point in the search space, paired with its corresponding objective function value, obtained through non-analytical means, e.g., sensors, experiments, or simulations. Common approaches include evolutionary optimization and other metaheuristics. Since BBO methods rely solely on objective function values, they typically require many evaluations, which becomes problematic when evaluating the ob- jective function is time-consuming or expensive. This leads to using surrogate-based optimization which evaluates selected true objective val- ues and trains a regression model to approximate the objective function across the search space. Surrogate-assisted black-box optimization is a small-data learning problem because the optimizer must approximate an expensive objective function from limited evaluations. Surrogate models act as data-efficient regressors, guiding the search toward promising or informative points under a restricted evaluation budget. In this paper, a new surrogate model using artificial neural networks, called Adaptive- Fidelity Nexus Covariance Matrix Adaptation Evolution Strategy (AFN- CMA-ES), is proposed for the selective evaluation of objective functions. The experimental results show its competitive performance compared to state-of-the-art surrogate-assisted BBO methods.

Visual graph reasoning requires answering graph-theoretic questions directly from graph images, where graph topology and state are conveyed visually rather than given in symbolic form. Despite recent progress of vision-language models (VLMs), current approaches to visual graph reasoning still fail on simple visual graph problems. This reveals a fundamental limitation of existing approaches: they prioritize final-answer supervision over the intermediate recovery of an explicit graph representation that preserves graph topology and state from visual input. To address this limitation, we propose VGCompiler, a compilation-centric paradigm for visual graph reasoning via knowledge compilation. VGCompiler organizes reasoning around two compilers: a representation compiler that recovers a structure-preserving intermediate graph representation from visual input, and an operation compiler that compiles query intent under the recovered graph state into an executable graph operation. Specifically, we build VGCompiler on Qwen3-VL-8B and train it with reinforcement learning guided by a layered reward over executability, compiled graph validity, representation quality, and operation quality. VGCompiler uses a frozen observer to summarize graph and question conditions into lightweight signatures, enabling archive retrieval and code reuse across similar regimes. Experiments on three benchmarks GVLQA, VisionGraph, and VGCURE, show that Qwen-VGCompiler, built on an 8B backbone, surpasses the strongest closed-source VLM baseline by 28.9% and the strongest code-based baseline by 23.7%, while maintaining high efficiency. We further evaluate VGCompiler on three real-world domains, including metro routing, logistics delivery, and network fault assessment, where it generalizes across heterogeneous visual graphs and domain-grounded tasks.

Business intelligence (BI) is a cornerstone of enterprise decision-making and is widely used by enterprise users in software such as Power BI and Tableau. In traditional BI workflows, users need to prepare data by (1) identifying relevant tables, (2) performing data transformations, and (3) building join relationships, before they can (4) answer their business questions. These steps can be complex and time-consuming, making BI challenging. Given the strong capabilities of large language models (LLMs) in working with data, we study their ability to answer BI questions end-to-end, without requiring users to manually perform the tedious preparation steps. To do this, we harvest a large collection of real-world BI projects from public sources, and manually extract pairs of (questions, ground-truth answers) from real user dashboards. The resulting benchmark, BI-Bench, is the first benchmark to systematically study LLMs' ability on end-to-end BI. We find that even frontier LLMs perform poorly on BI-Bench, with less than 50% accuracy. To address their limitations, we design a tool-augmented BI-Agent that decomposes BI workflows into subtasks on structured data, such as search, join, and transform, and orchestrates specialized data management methods across BI stages. Furthermore, we develop a post-training framework that synthesizes training trajectories from real BI projects, enabling BI-Agent to be further post-trained using both supervised fine-tuning (SFT) and reinforcement learning (RL). BI-Agent achieves substantial accuracy gains of up to 40 percentage points with vanilla LLMs, and post-trained BI-Agent yields gains of up to 30 points. Our results highlight the importance of combining tool-augmented reasoning with domain-specific post-training in complex BI workflows, and point to promising directions for future research.

Direct imaging of exoplanets is limited by the extreme contrast between the star and the planets, which is mitigated using a coronagraph. However, optical aberrations cause starlight leakage through the coronagraph, producing speckles that obscure the planetary signal. Achieving the required contrast levels demands wavefront control with subnanometric precision. Deep reinforcement learning offers a promising alternative to traditional focal-plane wavefront control techniques by enabling adaptive correction strategies learned directly from interaction with the system. In this work, we present a fully data-driven method for post-coronagraphic aberration correction in a simulated high-contrast imaging testbed. The agent controls a deformable mirror using observations consisting of focal-plane measurements (images) and physics-informed wavefront sensing information derived from these images. We evaluate different observation representations and control strategies, and the method is validated on simplified simulations of a high-contrast imaging testbed, where it successfully creates dark holes, i.e., regions of the focal plane in which residual starlight is strongly suppressed, while approaching the performance of conventional wavefront control methods.

As model-generated content is increasingly consumed by other model instances in automated workflows, a practically important question arises: can a model embed a signal in natural language that an independent instance of the same model can detect, relying only on shared pre-training and task instructions, without any shared memory or coordination-specific training? We introduce For Your Eyes Only, a cooperative signalling game designed to evaluate this directly. A Sender produces free-form descriptions for two words, one of which is a hidden target; an isolated Receiver must identify it. We evaluate seven contemporary models from four architectural families on 300 word pairs from established psycholinguistic corpora, using the Double-Pass Success Rate to control for output biases. We find that most models struggle to maintain coordination once they are required to avoid detectable signals, while one frontier model retains near-perfect performance even after such filtering. We further show that models can direct this capability toward deliberate misdirection, and that coordination is consistently weaker across architectures than within them.

Modern LLMs and AI agents increasingly support data engineering workflows that integrate evidence from unstructured sources. Such pipelines typically do data retrieval, integration, and ranking before proceeding to more complex agentic reasoning or actions, e.g., for scientific discovery. The core retrieval problem in these workflows jointly executes multi-attribute filtering, multi-vector search, exact relational joins, and thresholded embedding-similarity joins. Given a planned query and monotone scoring function, our DASE query engine constructs and ranks candidate evidence tuples. It comprises (i) a multi-step reasoning query model over structured predicates, multiple vectors, and relational links; (ii) SemJI, a sparse materialized embedding-similarity join index for rare near-neighbor pairs; and (iii) a co-designed execution layer that combines predicate-aware ANN traversal, batched access, and threshold-based score aggregation. On scientific-discovery workloads, DASE retrieves candidate evidence for multi-step reasoning queries 6x to 46x faster than strong RDBMS, rerank, and vector-database baselines at comparable recall; and for tasks that require semantic-operator post-processing, DASE acts as a high-recall prefilter that makes downstream LLM evaluation both cheaper and more accurate – e.g., on SemBench E-Commerce it improves BigQuery quality from 0.67 to 0.80 while cutting cost from $2.42 to $0.54.

Autonomous systems increasingly rely on Large Language Models (LLMs) yet the safety infrastructure surrounding these models introduces latency and compute overhead. This limits utility in resource-constrained, time-critical deployments. Existing external guardrail models remain blind to the model's internal workings, creating a fundamental assurance gap. We ask: does the model already know when the content is harmful? We extract activations from LLaMA-3.1-8B and train lightweight MLP classifier probes (12.6M parameters) to detect harmful prompts. Evaluated on WildJailbreak, Beavertails, and AEGIS 2.0, our probes achieve F1 scores of 99%, 83%, and 84%, respectively competitive with 1000x larger guard models while cutting latency and compute costs.

Compositional reasoning is critical for real-world problem solving: since training data is necessarily limited, models must generalize by composing learned skills in new ways. While post-training methods such as reinforcement learning (RL) have substantially improved the reasoning abilities of language models (LMs), their effects on compositional reasoning remain less well understood. We propose a dependency-graph framework to formalize compositional reasoning, yielding three levels of compositionality with increasing complexity. Empirically, we instantiate this framework with data-structure tasks, which provide deterministic reward computation and clear compositional structure. We find a consistent decomposed-to-composed asymmetry: decomposed-skill training does not reliably transfer to composed tasks, whereas composed-task training transfers more readily back to decomposed tasks. We provide theoretical explanation for this asymmetry, and further evaluate compositional generalization under length extrapolation, structural distribution shift, and transfer to tasks requiring unseen skills. Finally, we present a pilot study on real-world tool-calling benchmarks, showing preliminary evidence that the decomposed-to-composed asymmetry can extend to practical settings.

Retrieval-augmented systems increasingly rely on vector indexes that may retain deleted items in their search graph. Existing deletion interfaces can prevent deleted identifiers from appearing in returned results while still computing distances to their embeddings during graph traversal. We formalize this distinction as output safety versus traversal safety, and introduce TSD-AUDIT, a framework for auditing and enforcing traversal-safe deletion in graph-based approximate nearest-neighbor retrieval. On Faiss IndexHNSWFlat, native filtering leaves the number of distance computations unchanged relative to unfiltered search; at a 70% deletion rate, trace-faithful replay detects deleted-vector scoring in all 100 audited queries. Code inspection of hnswlib's mark_deleted path reveals the same scoring-before-liveness pattern. TSD-AUDIT enforces an alive-before-scoring invariant, repairs connectivity using only live candidates, and emits per-query scored-trace certificates that an independent verifier can check against the deletion snapshot. Under region-targeted deletion, TSD-AUDIT improves Recall@10 over native filtering by 4.3–42.2 percentage points across deletion fractions from 0.5 to 0.9, while remaining comparable under random deletion. These results show that output-only deletion audits can miss process-level exposure: auditing deletion in vector retrieval requires accounting for the vectors scored during search, not only the identifiers returned.

In both cognitive science and computer science, goals are conceptualized as cognitive states that flexibly combine with world knowledge to organize and specify purposeful behavior. In this way, goals are compositional representations whose content relates to rational behavior. We here draw attention to goals as representations and their content because it highlights a parallel with other areas in cognitive science - in particular, the syntax-semantics interface in linguistics and logic - while also foregrounding foundational questions about the expressivity, design, and efficiency of different goal representations. For example, goals are typically taken as fixed and imposing constraints on desirable behaviors, but we can also identify constraints on goal representations themselves, such as whether a particular goal language is sufficiently expressive to capture behaviors of interest, or whether different goal representations capture the same behavior. Here, we synthesize work that aims to characterize the properties of different goal representations and suggest these are points of a broader design space. We close by discussing how distinguishing the form and meaning of goals can elucidate the implicit assumptions we make about goals, inform the study of interactions between higher-level cognition and motivation, and isolate axes of variation for different conceptions of goals.

The rapid expansion of multimodal models has surfaced formidable bottlenecks in computation, memory, and deployment, catalyzing the rise of Efficient Multimodal Learning (EML) as a pivotal research frontier. Despite intensive progress, a cohesive understanding of what, how, and where efficiency is manifested across the learning stack remains fragmented. This survey systematizes the EML landscape by introducing the first structured, model-to-system taxonomy. We distill insights from over 300 seminal works into three hierarchical levels–model, algorithm, and system–addressing architectural parsimony, execution refinement, and hardware-aware orchestration, respectively. Moving beyond a purely categorical review, we offer a methodological synthesis of the vertical synergies between these layers, elucidating how cross-layer co-design contributes to the fundamental "Efficiency-Utility-Privacy" trade-off. Through an integrative case study of Multimodal Large Language Models (MLLMs), we trace the field's evolutionary trajectory from initial structural adjustments to modern full-stack resource orchestration. Furthermore, we provide a holistic discussion and application-specific optimization blueprints for diverse domains and posit a paradigm shift toward self-regulating intelligence, where efficiency is an intrinsic, emergent property of the model's fundamental design rather than a post-hoc constraint. Finally, we present open challenges and future directions that will define the trajectory of EML research. This survey establishes a structured framework for multimodal systems that are not only high-performing and generalizable but natively efficient and ready for ubiquitous deployment. A continuously updated version is available at https://github.com/pwang322/Efficient-Multimodal-Learning-Survey.

World action models (WAMs) rely on video-generation backbones, requiring substantial memory and compute for deployment. Post-training quantization reduces memory and can accelerate inference, but bit width, grouping, and quantizer choice define a large configuration space. Identifying configurations that preserve task performance through exhaustive closed-loop evaluation is costly. We propose PreDE (Predict Before You Deploy), a policy-calibrated framework for predicting quantization-induced task degradation from offline action deviations. Using closed-loop outcomes from a small development set, PreDE calibrates two thresholds and accepts, rejects, or defers new configurations using a fixed observation log. Under a within-setting label-ordering hypothesis, the rule issues decisions where all thresholds consistent with the development labels agree. Across five WAMs and four benchmark settings, quantization produces configuration-dependent task losses that cannot be explained by bit width alone or a shared deviation threshold. Across 28 held-out configurations from two policies, PreDE issued 21 decisions before observing closed-loop outcomes (75% coverage), all matching the observed acceptable or degraded labels. Deferred candidates included both acceptable outcomes and a 33-percentage-point loss. In 450 Franka Research 3 trials across two independently fine-tuned policies, all configurations assigned to high-deviation groups before testing showed significant degradation, while low-deviation comparisons showed no statistically significant degradation. On the real robot, W4A4 achieved a 1.37x action-query speedup and approximately 44% lower peak memory. These results support policy-specific behavioral calibration for quantization configuration selection while identifying candidates that require closed-loop evaluation. The code is available at https://github.com/jiuyixu25/PreDE.

Tool-augmented large language model (LLM) agents fail in a way no tool-selection or tool-security method addresses: they call tools that do not exist and pass arguments no schema declares. Existing defenses either pick the right tool (selection) or constrain what an agent may do with real tools (gating), both of which presuppose the emitted call refers to a real tool at all. We show this is a structural blind spot: a hallucinated call is by construction not a decision any gate made, so no gate can reject it. This paper is primarily a measurement and benchmark study. We give a five-class taxonomy of tool hallucination (H1-H5) and, as a reference point, the Resolution Rung: a training-free, closed-world resolver (registry membership plus a signature check) whose interest is where it must sit, not what it computes. We prove hallucination defense must precede any causal gate, and characterize the one irreducible residue (borrowed arguments schema-indistinguishable from a valid call). Across ten hosted models under two invocation surfaces we measure 322 genuine hallucinations; fabricated-tool calls concentrate on the unconstrained raw-JSON surface (34 vs. 3), and model scale does not help (a 675B model matches a 7-8B one). We then extend to the Model Context Protocol, where merging several servers into one namespace creates hallucination surfaces a single registry cannot express (a second taxonomy, M1-M5); on the live MCP surface we measure 154 hallucinations, including from frontier models that were clean on the single-registry surface, because collisions and shadowing are structural to the merge. We release the versioned Hallucinated-Tools Benchmark (HTB) so any resolver is comparable across submissions.

Robot manipulation policies are improving quickly, and real-robot evaluation remains the standard evidence for that progress. It still relies on a human to reset the scene between rollouts, which consumes operator time and leaves the initial state distribution unspecified, so results reproduce poorly. A recent system, AutoEval, automates both reset and scoring, but only for single-step tasks, because a long-horizon rollout can terminate in combinatorially many configurations that no single learned reset policy covers. We present HALTER, a Harness for Autonomous Long-horizon Task Evaluation and Reset, which restores the scene by planning over a library of learned atomic reset skills, so demonstration cost scales with the size of that library rather than with the number of terminal states. HALTER builds a spatial scene graph online from point clouds and vision foundation models, and an LLM reasons over this graph to score the rollout, plan the reset, and verify that the reset succeeded, without collecting labeled success images for any task. On four long-horizon tasks on a Franka arm, HALTER restores the scene in 76% of episodes, against 52% for AutoEval and 65% for a motion-planning reset, and it estimates the completed-skill fraction correctly in 90% of episodes, against 76%. Its reset-verification verdict is correct in 91% of episodes, compared with 78% for AutoEval. It also cuts the operator time of an evaluation campaign by 72% relative to manual reset. We further measure compositional generalization on three held-out tasks, where HALTER resets 74.7% of episodes against 1.3% for a per-task reset policy, and we ablate the scene representation and the graph update rate.

Game-theoretic analyses of cyber defence often compute equilibria of games whose payoffs exist only as the output of a simulator. Iterative equilibrium-finding methods grow a set of attacker and defender policies and need the payoff of every attacker–defender pair, so they are bottlenecked by payoff estimation: each payoff costs many simulator runs. We introduce Regret-Weighted Payoff Sampling (RWPS), which spends a fixed simulation budget on the payoffs the equilibrium actually depends on and predicts the rest with a model trained on every payoff measured so far. Standard error bounds for estimated games are driven by the worst-estimated payoff, so they cannot credit an estimator that is inaccurate only where accuracy does not matter. We prove a bound that weights payoff errors by the opponent's equilibrium strategy, a certificate that can be computed from simulated payoffs alone, and a condition under which errors in the predicted payoffs cannot change either player's regret. On three synthetic general-sum games, one of them a Colonel Blotto game of military resource allocation, the new bounds are four to six times tighter than the standard one, and RWPS finds less exploitable equilibria than minimum-regret-first search, information-gain search and progressive sampling at the same budget. On two cyber-defence simulators, CyGym and a new game whose hosts are LLM agents exposed to prompt injection, it gives the least exploitable equilibria at the smallest budgets.