Academic Intelligence · Curated Daily

Explore the Frontier of Global Academia

AcademicHub aggregates real-time literature from top journals and preprint platforms. Build your personal research radar and let large language models compile cross-disciplinary analysis briefings automatically.

01.
arXiv (CS.CL) 2026-06-15

Rethinking the Trust Region in LLM Reinforcement Learning

Reinforcement learning (RL) has become a cornerstone for fine-tuning Large Language Models (LLMs), with Proximal Policy Optimization (PPO) serving as the de facto standard algorithm. Despite its ubiquity, we argue that the core ratio clipping mechanism in PPO is structurally ill-suited for the large vocabularies inherent to LLMs. PPO constrains policy updates based on the probability ratio of sampled tokens, which serves as a noisy single-sample Monte Carlo estimate of the true policy divergence. This creates a sub-optimal learning dynamic: updates to low-probability tokens are aggressively over-penalized, while potentially catastrophic shifts in high-probability tokens are under-constrained, leading to training inefficiency and instability. To address this, we propose Divergence Proximal Policy Optimization (DPPO), which substitutes heuristic clipping with a more principled constraint based on a direct estimate of policy divergence (e.g., Total Variation or KL). To avoid huge memory footprint, we introduce the efficient Binary and Top-K approximations to capture the essential divergence with negligible overhead. Extensive empirical evaluations demonstrate that DPPO achieves superior training stability and efficiency compared to existing methods, offering a more robust foundation for RL-based LLM fine-tuning. Our code is available at https://github.com/sail-sg/Stable-RL.

02.
arXiv (CS.CL) 2026-06-16

Not What, But How: A Framework for Auditing LLM Responses across Positioning, Generalization, Anthropomorphism, and Maxims

Large language models (LLMs) are being increasingly used to answer subjective, information-seeking questions, where users are sensitive to how responses are communicated, not just whether the answers are correct. Existing LLM evaluations for subjective cultural queries largely focus on factual correctness, ignoring how the response is framed. To this end, we introduce FRANZ, an automated FRAmework for respoNse characteriZation to conduct communicative audit of LLM responses along four dimensions: cultural positioning, use of generalizing language, anthropomorphic cues, and adherence to conversational maxims. To enable this evaluation, we contribute SQUARE - a corpus of 376k subjective questions sourced from 57 subreddits, and mapped to 7 countries and 19 question categories. We demonstrate FRANZ's applicability by scoring responses from three open-weight LLMs. We observe that LLMs show statistically significant differences in the frequency with which they employ each response characteristic. Unlike single-dimensional audits, FRANZ reveals that insider positioning and anthropomorphism are positively coupled, with the degree of coupling varying by country, providing a diagnostic lens for identifying framing divergences.

03.
arXiv (CS.CL) 2026-06-15

GitOfThoughts: Version-Controlled Reasoning and Agent Memory You Can Replay, Diff, and Merge

Large language model (LLM) reasoning is ephemeral: chains of thought vanish with the context window, pruned search branches leave no record, and memory buffers cannot be diffed, merged, or audited. Every other complex software process (code, infrastructure, data, experiments) is version-controlled; reasoning is not. We introduce GitOfThoughts, which stores an agent's reasoning tree as a git repository: every scored thought is a commit, scores are notes, outcomes are tags, and retrieval is "git log" over the agent's own history. This makes reasoning replayable, auditable, and mergeable across agents at near-zero engineering cost. We then ask the harder question: does memory, in any substrate, actually improve accuracy? Across five substrates (none, markdown, vector, graph, git), two benchmarks, two model scales, and pre-registered replications, the answer for novel problems is no. No memory format reliably helps, and a promising early result collapsed under its own pre-registered replication. Memory pays only above what we call the copyability threshold: when the retrieved case is a near-duplicate of the current problem (similarity >~ 0.8), accuracy jumps sharply; below it, nothing. The gain is answer retrieval, not method transfer: a 4.5x larger model doubles the near-duplicate payoff yet still cannot extract a transferable method from a worked example. The only general lever we find is test-time sampling. The case for git-as-substrate is therefore auditability, provenance, and mergeability at accuracy parity. We document a retracted result and a refuted hypothesis to model the evaluation standard we hold ourselves to.

04.
medRxiv (Medicine) 2026-06-22

ECG-Guided Pre-Screening of Family Members for Hypertrophic Cardiomyopathy

Background: Current clinical guidelines recommend serial ECG and echocardiographic surveillance for first-degree relatives of probands with Hypertrophic Cardiomyopathy (HCM). Objectives: To evaluate the accuracy and validity of ECG alone as a pre-screening tool for the diagnosis of HCM and to develop a random forest (RF) model for HCM phenotype prediction. Method: Pediatric relatives of primary HCM probands attending the cardiomyopathy screening program at The Hospital for Sick Children were included from 1993 to 2025. Subjects were followed until the last follow-up, censored at phenotype conversion. ECGs were classified as normal or abnormal based on predefined parameters. Associations between binary ECG variables and HCM phenotype were assessed using Phi ({varphi}) coefficient. A Random Forest classifier was developed using significant ECG variables (70:30 training: test split) and evaluated using precision, recall, specificity, negative predictive value, F1 score and AUROC. Feature importance was assessed using SHAP analysis. Variables with an impact of >5% were included in a simplified model, which was evaluated by repeating performance metrics and externally validated in a healthy cohort. Results: 350 screened relatives (44% female, mean follow-up 6.8 +- 4.8 years) were included. At baseline, 13% (46350) were phenotype-positive for HCM. 9 subjects converted during the surveillance. Thirteen ECG variables were significantly associated with phenotype-positive HCM and were included in the full random forest model. Four variables had >5% impact (Left ventricular hypertrophy, right ventricular hypertrophy, T-wave inversion and ST-segment depression) and were included in a simplified model, which maintained high specificity (93% vs 97%), negative predictive value (97% vs 93%) and AUROC (90% vs 96%). The simplified model classified 83% subjects as phenotype-negative, with eight being false-negative, all of whom developed an abnormal ECG in a mean of 1 year, and none had an interim adverse cardiac event. The simplified model was evaluated in an independent healthy cohort of 153 school-age subjects and correctly identified 98% as phenotype-negative with 100% NPV. Conclusion: ECG abnormalities were strongly associated with phenotype-positive status. A simplified ECG-based random forest model using four ECG variables demonstrated high specificity and negative predictive value for identifying phenotype-negative subjects. If prospectively validated, this could reduce the need for concurrent echocardiographic screening by up to 83% per encounter, lowering screening burden and cost.

05.
arXiv (CS.CV) 2026-06-11

Weakly Supervised Segmentation as Semantic-Based Regularization

Weakly supervised semantic segmentation (WSSS) trains dense pixel-level segmentation models from partial or coarse annotations such as bounding boxes, scribbles, or image-level tags. While recent work leverages foundation models such as the Segment Anything Model (SAM) to generate pseudo-labels, these approaches typically depend on heuristic prompt choices and offer limited ways to incorporate prior knowledge or heterogeneous labels. We address this gap by taking a neurosymbolic perspective: integrating differentiable fuzzy logic with deep segmentation models. Weak annotations and domain-specific priors are unified as continuous logical constraints that fine-tune SAM under weak supervision. The refined foundation model then produces improved pseudo-labels, from which we train a second-stage prompt-free segmentation model. Experiments on Pascal VOC 2012 and the REFUGE2 optic disc/cup segmentation dataset show that our logic-guided fine-tuning yields higher-quality pseudo-labels, leading to state-of-the-art segmentation accuracy that often exceeds densely supervised baselines.

06.
arXiv (CS.LG) 2026-06-17

A Closer Look at Failure Modes in Temporal Understanding of Large Audio-Language Models

arXiv:2606.17417v1 Announce Type: cross Abstract: Large Audio Language Models (LALMs) achieve strong performance on a variety of audio understanding tasks but continue to struggle with temporal reasoning, a fundamental capability central to human auditory perception. Understanding the causes of these failures remains challenging as existing benchmarks report performance gaps without probing underlying mechanisms. To address this, we introduce a benchmark with 1,657 questions across three foundational tasks designed specifically for mechanistic analysis. Examining model outputs across varying input settings (behavioral analysis) reveals that models often under-utilize audio when textual cues are available. We also provide the first causal mechanistic analysis of temporal reasoning failures in LALMs. Comparing attention upweighting against scaling, we find that redistributing attention across audio tokens is more effective than increasing audio attention. Targeting task-relevant tokens yields further gains. These findings suggest that modality imbalance alone cannot explain failures. Attention scaling at bottleneck layers improves accuracy from 55.9% to 59.1% without fine-tuning, demonstrating a promising direction for future work.

07.
arXiv (quant-ph) 2026-06-19

Effective Faraday interaction between light and Helium-3 nuclear spins in a multi-pass cell

arXiv:2606.20328v1 Announce Type: new Abstract: Helium-3 nuclear spins form an exceptionally stable quantum system with extremely long coherence time, offering exciting opportunities for quantum technologies. In particular, nuclear spin-squeezed states promise enhanced precision for sensing tasks and tests of new physics. A central challenge for all these applications is the realization of a controllable light-nuclear spin interface. Here we experimentally demonstrate such an interface by exploiting metastability-exchange collisions in a low-pressure helium-3 gas cell at room temperature. A radio-frequency discharge produces a small population of metastable atoms that both enables efficient optical pumping and mediates an effective Faraday interaction between the collective nuclear spin and an optical probe. We quantitatively characterize the strength of this interaction as a function of the nuclear polarization, applied magnetic field, and probe-beam parameters. Moreover, we show that using a multi-pass cell enhances this interaction by effectively increasing the optical depth. Extrapolating to a tenfold increase of the probe power used in the present experiment, we project a measurement-induced squeezing rate of 0.52 s$^{-1}$. Our results provide a practical pathway for optical access to helium-3 nuclear spins and open prospects for generating long-lived, macroscopic nuclear spin-squeezed states for quantum metrology.

08.
arXiv (quant-ph) 2026-06-16

Intrinsic preservation of plasticity in continual quantum learning

arXiv:2511.17228v2 Announce Type: replace Abstract: Artificial intelligence in dynamic, real-world environments requires the capacity for continual learning. However, standard deep learning suffers from a fundamental issue: loss of plasticity, in which networks gradually lose their ability to learn from new data. Here we show that quantum learning models naturally overcome this limitation, preserving plasticity over long timescales. We demonstrate this advantage systematically across a broad spectrum of tasks from multiple learning paradigms, including supervised learning and reinforcement learning, and diverse data modalities, from classical high-dimensional images to quantum-native datasets. Although classical models exhibit performance degradation correlated with unbounded weight and gradient growth, quantum neural networks maintain consistent learning capabilities regardless of the data or task. We identify the origin of the advantage as the intrinsic physical constraints of quantum models. Unlike classical networks where unbounded weight growth leads to landscape ruggedness or saturation, the unitary constraints confine the optimization to a compact manifold. Our results suggest that the utility of quantum computing in machine learning extends beyond potential speedups, offering a robust pathway for building adaptive artificial intelligence and lifelong learners.

09.
arXiv (CS.CL) 2026-06-12

Unraveling Syntax: Language Modeling and the Substructure of Grammars

While language models achieve impressive results, their learning dynamics are far from understood. Many domains of interest – such as natural language syntax, coding languages, arithmetic – are captured by context-free grammars (CFGs). In this work, we extend prior work on neural language modeling of CFGs in a novel direction: how language modeling behaves with respect to CFG substructure, namely subgrammars. We define subgrammars, and prove a set of fundamental theorems connecting language modeling and subgrammars. We show that language modeling loss recurses linearly over its top-level subgrammars; applied recursively, the loss decomposes into losses for "irreducible" subgrammars. Under additional assumptions, and empirically, parametrized models learn subgrammars in parallel, unlike children who first master simple substructures. We find that subgrammar pretraining can improve final performance, but only for tiny models relative to the grammar, while alignment analyses show that pretraining consistently leads to internal representations that better reflect the grammar's substructure.

10.
arXiv (quant-ph) 2026-06-12

Unifying spacetime approaches to quantum mechanics

arXiv:2606.12539v1 Announce Type: new Abstract: Recent efforts to formulate quantum mechanics in a way that treats space and time on a more equal footing have led to a large variety of spacetime-oriented approaches. In this work we present a detailed study of spacetime states, the objects that play the role of quantum states in the recently introduced framework of spacetime quantum mechanics, and show that the main proposals in the literature are different manifestations of the same underlying object. Path integrals, quantum states over time, pseudo-density matrices, the Page and Wootters mechanism, superdensity operators, and timelike-entanglement proposals all arise from spacetime states through particular evaluations, reduced information, linear maps, or quantum channels. This unification provides explicit mathematical representations of these formalisms, reveals relations among them, and clarifies the spacetime information each one captures. We also study the broader relevance of the spacetime-state point of view for Leggett-Garg inequalities, OTOCs, temporal tensor networks, fermionic systems, relativistic QFTs, quantum reference frames, and classical physics, together with additional insights and perspectives revealed by the common unifying framework.

11.
arXiv (CS.CV) 2026-06-24

Systematic Exploration of 4-Expert Heterogeneous Mixture-of-Experts via Automated Pipeline Search

We present an automated large-scale search pipeline for heterogeneous 4-Expert Mixture-of-Experts (MoE4) architectures within the LEMUR neural network dataset ecosystem. Building on a hand-crafted heterogeneous MoE reference model, we replace manual design with a deterministic code-assembly generator that systematically combines base architecture families drawn from the LEMUR database into MoE4 ensembles, each governed by a convolutional gating network with temperature scaling, mixup augmentation, and cosine-annealed learning rate scheduling. Over a 28-day campaign on an NVIDIA RTX 4090, the pipeline generated 4,463 candidate models across 197 batches, of which 1,021 were evaluated successfully. A critical finding emerged from the campaign: due to alphabetical enumeration via itertools.combinations, the entire explored search space (4.8% of the theoretical 23,751 possible 4-family combinations) is anchored to a single family, AirNet. We characterise this coverage bias precisely, identify the root cause in the generator, and propose a stratified random sampling fix. Within the AirNet anchored scope, ShuffleNet and MobileNetV3 consistently co-produce the highest-accuracy ensembles (mean accuracy up to 0.632), while FractalNet and MNASNet are identified as low-yield families warranting exclusion in future campaigns. The pipeline, analysis artefacts, and corrected generator are released as part of the open-source NNGPT project at https://github.com/ABrain-One/nn-gpt

12.
arXiv (quant-ph) 2026-06-15

Certification of the genuine resolution of photon number resolving detectors

arXiv:2606.14365v1 Announce Type: new Abstract: Photon-number-resolving (PNR) detectors are essential components of photonic quantum technologies, yet thus far, no practical metric exists to certify how many photons they can genuinely resolve in a single measurement. Here we introduce an operational framework for quantifying the capability of a PNR detector to distinguish between different numbers of photons, i.e. its genuine resolution. In turn, we develop a practical and scalable protocol for certifying the genuine resolution of a detector, which is based on coherent state probes. We apply the method to a 28-pixel photon-number-resolving superconducting nanowire single-photon detector (PNR-SNSPD) and certify genuine four-outcome resolution. Our work highlights the critical requirements in terms of detector efficiency towards achieving high genuine resolution. This approach provides an operational benchmark for PNR detectors and fills a crucial gap in the characterization of photonic quantum devices.

13.
arXiv (CS.CV) 2026-06-12

Heterogeneous LiDAR Early Fusion and Learned Re-Ranking Strategy for Robust Long-Term Place Recognition in Unstructured Environments

Robust localization in unstructured environments, such as agricultural fields, is a critical challenge for autonomous systems. LiDAR sensors provide detailed 3D information about the environment and are invariant to lighting conditions. For this reason, LiDAR-based place recognition methods have gained significant attention. In this paper, we propose MinkUNeXt-VINE++, a novel approach that combines early fusion of heterogeneous LiDAR data from two sensors (Livox Mid-360 and Velodyne VLP-16) and a learned re-ranking strategy in inference time. This fusion leverages the strengths of each sensor to provide a more comprehensive representation of the environment. Additionally, the re-ranking approach is particularly important in repetitive environments, such as vineyards, as finding true positives is a major challenge. We evaluated our approach using the TEMPO-VINE dataset, which provides heterogeneous LiDAR data in vineyard environments across different phenological stages. Our results demonstrate that MinkUNeXt-VINE++ significantly improves place recognition performance compared to single-sensor approaches and state-of-the-art methods. MinkUNeXt-VINE++ achieves a 20% improvement in the Recall@1 metric compared to single-sensor approaches, and +30% including re-ranking. The code of our method is publicly available for reproduction.

14.
arXiv (CS.AI) 2026-06-24

A Survey on Federated Causal Discovery and Inference

arXiv:2606.23741v1 Announce Type: cross Abstract: Causal reasoning, which encompasses the discovery of causal structures and the inference of causal effects, is fundamental to data-driven decision making. In practice, data for reliable causal analysis are often distributed across institutions and cannot be centralized due to privacy regulations or communication constraints. Federated learning (FL) addresses this by enabling collaborative analysis without raw data sharing, giving rise to the rapidly growing field of federated causal discovery (FCD) and inference (FCI). However, the interdisciplinary nature of this field and the absence of a comprehensive survey present barriers to entry for researchers. This paper bridges that gap by providing a systematic review through multi-dimensional taxonomies. Grounded in the three core design decisions underlying any FCD solution, namely how structures are learned, how data are partitioned, and what structural knowledge each party obtains, we organize FCD along three axes: methodological paradigm, federation topology, and structural scope. We further examine key practical dimensions, including temporal dynamics, data heterogeneity, missing data, and non-identical variable sets. For FCI, we categorize methods by target estimand (average versus individualized/conditional treatment effects) and by estimation strategy, from classical weighting methods to modern deep generative architectures. Unlike prior works that treat FCD and FCI separately, we formalize their connection as complementary stages of a unified federated causal reasoning pipeline, where FCD supplies the structural knowledge required for valid effect estimation in FCI. Finally, we highlight their shared concerns regarding privacy, communication efficiency, theoretical guarantees, and application domains, and conclude by identifying open challenges for future research.

15.
arXiv (CS.CV) 2026-06-25

Auto-Labelling-Based Domain Transfer for 3D Object Detection on a Bicycle-Mounted LiDAR Platform

Reliable 3D perception of vulnerable road users (VRUs) such as cyclists and pedestrians is essential for their safety in urban traffic and a core requirement for autonomous driving (AD). Alongside advances in vehicle-based perception, research increasingly equips bicycles with sensors to study traffic from a perspective native to VRUs. Such platforms still rely on LiDAR detectors originally trained on vehicle data, yet annotated 3D data from a cyclist's perspective is scarce. How well these detectors generalise to this setting has not been evaluated. We present a 3D object detection benchmark of 1,027 annotated LiDAR keyframes (over 18,000 3D bounding boxes) from the FUSE-Bike platform in urban Munich. We evaluate four nuScenes-pre-trained detectors against 1,854 human-verified ground-truth (GT) boxes both in their original form and after finetuning on training labels produced by a VRU-dedicated auto-labelling pipeline that requires no manual annotation. The zero-shot domain gap is concentrated on the VRU classes. Finetuning recovers most of it, improving mean average precision (mAP) by up to 23.4 points with the largest gains on pedestrians and cyclists, and the adapted detectors even surpass the quality of the auto-labels they were trained on. The benchmark provides a reproducible baseline for VRU-centric 3D detection and shows that auto-labels are a viable substitute for manual annotation when adapting vehicle-trained detectors to a cyclist platform.

16.
arXiv (CS.LG) 2026-06-17

Amortizing Maximum Inner Product Search with Learned Support Functions

arXiv:2603.08001v2 Announce Type: replace Abstract: Maximum inner product search (MIPS) is a crucial subroutine in machine learning, requiring the identification of a vector taken within a database (the keys) that best aligns with a given query. We propose amortized MIPS: a regression-based approach that trains neural networks to directly predict MIPS solutions, amortizing the cost of repeatedly solving MIPS for queries drawn from a known distribution over a fixed key database. Our key insight is that the MIPS value function is the support function of the set of keys, a well-studied convex function whose gradient yields the optimal key. This motivates two complementary amortized models: SupportNet, an input-convex neural network trained to regress the support function, and KeyNet, a vector-valued network that directly regresses the optimal key. SupportNet can serve as a cluster router, steering queries toward relevant database partitions, while KeyNet can be used as a drop-in replacement for the original query, fed directly to off-the-shelf indexing pipelines. Our experiments on the BEIR benchmark show that, for document embeddings, learned \SupportNet{}s and \KeyNet{}s significantly improve IVF match rates when accounting for compute effort, whether measured in FLOPs, number of probes, or wall-clock time. Our code is available at: https://github.com/apple/ml-amips.

17.
arXiv (CS.AI) 2026-06-18

CAPRA: Scaling Feedback on Software Architecture Deliverables with a Multi-Agent LLM System

arXiv:2606.18976v1 Announce Type: cross Abstract: Automated assessment in software engineering education has advanced significantly for code grading and essay scoring. However, reviewing software architecture deliverables, which requires analyzing structural completeness and requirements traceability, has not yet been fully automated. Applying Large Language Models (LLMs) to this task requires robust architectures to ensure technical feedback is accurate and reliable for students. This paper presents CAPRA (Configurable Architecture Proficiency Report Assessment), a multi-agent LLM system that analyzes software architecture deliverables to generate personalized, template-compliant LaTeX feedback. As a core design choice, CAPRA coordinates multiple specialized agents and employs a Python-based microservice for multi-modal document extraction, utilizing PyMuPDF and vision-enabled LLMs (specifically gpt-4o) to parse text and UML diagrams. To ensure educational reliability and mitigate hallucinations, CAPRA introduces a deterministic Evidence Anchoring step using fuzzy matching via normalized Levenshtein distance, along with a ConsistencyManager agent that cross-verifies, deduplicates, and merges findings. System performance is assessed using a structured eight-criterion binary evaluation taxonomy covering: (i) extraction completeness, (ii) feature validation, (iii) issue grounding and severity detection, (iv) recommendation specificity and traceability, and (v) template and tone compliance. A preliminary empirical evaluation on 10 student reports shows that CAPRA satisfied 88.8% of the evaluated criteria under a strict two-rater aggregation rule, achieved moderate inter-rater agreement with human evaluators (kappa = 0.582), and processed each report in slightly over 4 minutes. While these results support the viability of LLM-supported architectural feedback, human oversight remains essential for subjective assessment dimensions.

18.
arXiv (CS.AI) 2026-06-11

When Does Deep RL Beat Calibrated Baselines? A Benchmark Study on Adaptive Resource Control

arXiv:2605.26418v2 Announce Type: replace-cross Abstract: A properly calibrated rule-based autoscaler can beat every one of six mainstream deep reinforcement learning (DRL) algorithms on cost across every workload we test - so when, if ever, does DRL actually help? We study this in RLScale-Bench, a reproducible benchmark and evaluation protocol for DRL on adaptive resource control, where an agent allocates compute to a dynamic workload under cost and service-level constraints. We evaluate PPO, DQN, A2C, SAC, TD3, and DDPG under matched architectures, training budgets, and reward functions against a calibrated rule-based baseline across six workload patterns and five seeds (240 runs), instantiate the benchmark on Kubernetes Horizontal Pod Autoscaling, and probe distribution-shift generalization. Three findings challenge common assumptions: (i) the calibrated controller achieves the lowest cost on all six workloads, though it trails the best RL agents on bursty and flash traffic; (ii) discrete-action algorithms outperform continuous-action ones by one to two orders of magnitude in constraint violations due to action-space mismatch; and (iii) no single algorithm dominates across workloads, with rankings shifting by up to four positions. The bottleneck in RL-based resource control is not algorithm selection but baseline calibration, reward engineering, and realistic evaluation protocols.

19.
arXiv (CS.CV) 2026-06-25

Neural Network Quantization by Learning Low-Loss Subspaces

Neural network quantization aims to find a discrete representation of parameters that preserves the performance of a full-precision (FP) model as faithfully as possible. Enforcing discrete constraints perturbs parameters away from a well-optimized minimum, generally resulting in performance degradation. Recent studies indicate that low-loss FP solutions are not isolated, but instead belong to connected low-loss subspaces of the loss landscape, where the loss maintains nearly the same minimum value. Models sampled from these subspaces are diverse and retain high accuracy. This raises the question: can a quantized model be constructed to lie within a low-loss subspace of the FP model, thereby automatically preserving performance? We address this question by learning quantization-aware linear paths in weight space optimized to minimize loss. We demonstrate that the midpoint of the resulting subspace is, by design, quantization-friendly and that its direct quantization yields performance comparable to that of quantization-aware training. The proposed procedure offers a novel perspective on weight quantization and, in contrast to conventional methods, neither relies on the straight-through estimator nor involves explicit discretization during training.

20.
arXiv (CS.AI) 2026-06-19

Reward as An Agent for Embodied World Models

arXiv:2606.19990v1 Announce Type: new Abstract: While RL has become a promising tool for refining world models, existing methods largely rely on conservative rollouts near the training distribution, limiting exploration, behavioral diversity, and richer dynamic discovery. In this work, we challenge this conservative paradigm. We argue that the core limitation is not exploration itself, but the lack of reliable verification strategies to support broader exploration. Without reliable verification, expanded exploration becomes highly susceptible to reward hacking, where policies exploit imperfect rewards without achieving genuine improvement. To evaluate this motivation, we instantiate our method in embodied world models, where physical plausibility, and task completion provide a rigorous testbed for scalable RL under complex dynamics. On the verification side, we introduce Reward as an Agent, an agentic reward framework that actively evaluates generated behaviors to provide robust reward signals and mitigate reward hacking under distribution shifts. On the exploration side, we introduce Dynamic-Aware Rollout Diversification through DynDiff-GRPO, which explicitly expands action-space exploration to diversify trajectories, broaden state-action coverage, and encourage richer embodied behaviors beyond conservative rollout regimes. By unifying Reward as an Agent with DynDiff-GRPO, we enable RL on a more reliable reward foundation with substantially diversified sampling, effectively mitigating reward hacking while yielding significant accuracy gains across multiple open-source world models, thereby demonstrating that broader exploration can scale successfully when grounded in robust verification.

21.
arXiv (CS.AI) 2026-06-16

Action with Visual Primitives

arXiv:2605.22183v3 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models have emerged as a promising paradigm for generalist robotic manipulation. A common design in current architectures maps language instructions and visual observations to actions in a single forward pass. While conceptually simple, this formulation entangles instruction comprehension, spatial scene understanding, and motor control within a single learning objective. As a result, the action expert must implicitly relearn cognitive and perceptual capabilities already present in the pretrained VLM, which can limit both learning efficiency and generalization. We introduce AVP (Action with Visual Primitives), an end-to-end architecture that implements this visual-primitive-centric interface: the VLM infers the next-stage target and emits visual-primitive tokens that condition a flow-matching action expert, with supervision derived from end-effector kinematics. Real-robot experiments on general pick-and-place tasks show that AVP improves the success rate by 37.04% over pi_0.5 and outperforms other recent methods, with consistent gains in data efficiency, spatial-compositional generalization, and object-level transfer.

22.
arXiv (CS.AI) 2026-06-25

SoK: AI Secure Code Generation: Progress, Pitfalls, and Paths Forward

arXiv:2606.25195v1 Announce Type: cross Abstract: The increasing use of AI systems for code generation raises a central security question: what can today's models and coding agents actually do to produce secure code, where do they still fail, and what would move the field forward? Existing work has explored prompting, fine-tuning, reinforcement learning, and agentic workflows for secure code generation, but the field still lacks a systematic understanding of how these techniques improve security and why substantial failures persist. In this SoK, we systematize the progress, pitfalls, and paths forward for AI secure code generation. We introduce a three-level framework that measures models' natural-language understanding of secure coding principles, their code-level actuation of those principles during generation, and the knowledge–actuation gaps between the two. We instantiate this framework across models and coding agents on benchmarks covering both isolated function-level security and full web-application security. Our results show that secure-coding-principle understanding is a statistically strong predictor of code-level outcomes, including functional correctness, security, and joint functional-security correctness. Yet substantial knowledge–actuation gaps remain: models can recognize relevant security principles but still fail to translate them into secure and functional code. These findings offer a principle-centered account of where AI secure code generation stands today and identify concrete paths forward through principle-guided generation, evaluation, benchmarking, and agentic workflows.

23.
arXiv (CS.AI) 2026-06-18

Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection

arXiv:2606.19168v1 Announce Type: new Abstract: To achieve deeper safety alignment for large language models (LLMs), recent efforts have studied how to push safety interventions earlier into the pretraining stage, primarily by filtering unsafe data or rewriting it into safer forms. We argue that pretraining-stage alignment should go beyond making the data safe: LLMs may compose seemingly benign knowledge and capabilities into unsafe behaviors. To this end, we propose Safety Reflection Pretraining, a pretraining-stage alignment method which regularly inserts short safety reflections into pretraining corpora to integrate self-monitoring directly into language modeling, establishing a foundational capability that is subsequently reinforced by compatible post-training. Our experiments with 1.7B models pretrained on FineWeb-Edu show that Safety Reflection Pretraining improves safety classification accuracy and substantially reduces the success rates of inference-stage and finetuning attacks. Complementary to our real-world experiments, we also introduce a fully controlled synthetic environment, MedSafetyWorld, with a clear definition of safety and a reasoning structure under which models can easily generalize unsafe behaviors from safe data. Ablations in MedSafetyWorld further demonstrate a clear advantage of Safety Reflection Pretraining in preventing models from acting on unsafe behaviors generalized from safe data, compared with data filtering and rewriting. Taken together, our findings suggest that pretraining alignment should not only make the training data safe, but also shape the behaviors that models are likely to acquire from safe data.

24.
arXiv (CS.AI) 2026-06-11

Position: Stop Anthropomorphizing Intermediate Tokens as Reasoning/Thinking Traces!

arXiv:2504.09762v4 Announce Type: replace Abstract: Intermediate token generation (ITG), where a model produces output before the solution, has become a standard method to improve the performance of language models on reasoning tasks. These intermediate tokens have been called \say{reasoning traces} or even \say{thinking traces} – implicitly anthropomorphizing the traces, and implying that these traces resemble steps a human might take when solving a challenging problem, and as such can provide an interpretable window into the operation of the model's thinking process to the end user. In this position paper, we present evidence that this anthropomorphization isn't a harmless metaphor, and instead is quite dangerous – it confuses the nature of these models and how to use them effectively, and leads to questionable research. We call on the community to avoid such anthropomorphization of intermediate tokens.

25.
arXiv (CS.CV) 2026-06-25

TACO: Towards Task-Consistent Open-Vocabulary Adaptation in Video Recognition

Adapting CLIP for open-vocabulary video recognition necessitates a delicate balance between newly acquired video knowledge and the pretrained generalization. While existing studies pursue this generalization-specialization trade-off with additional regularizations or constraints, we argue that they overlook the deviation of representations beyond the fine-tuning data distribution, resulting in suboptimal adaptation effects. We believe such deviation is inherited from the inconsistency between the fine-tuning and evaluation objectives, where model optimization is restricted to the known training distribution but evaluated on unseen ones. In this paper, we introduce TACO, a simple yet effective framework to mitigate the potential negative effects induced by this inconsistency. Our key insight is that adaptation should preserve OOD-relevant alignment beyond the training distribution. To this end, we propose Relative Structure Distillation, which regularizes the relative geometry of the representation space and suppresses harmful alignment shift during training. We further decouple the representation space from the optimization space with a lightweight specialization projection, allowing task-specific adaptation without directly overspecializing the representations used at test time. TACO establishes state-of-the-art performance on diverse benchmarks under cross-dataset and base-to-novel settings. Code will be released at https://github.com/ZMHH-H/TACO.