Academic Intelligence · Curated Daily

探索全球前沿学术脉络

AcademicHub 汇聚顶级期刊与预印本平台的实时文献。定制您的专属科研雷达,利用大语言模型自动生成交叉领域文献分析简报。

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

MooMIns – Monocular 3D Reconstruction and Object Pose Estimation from Multiple Instances

Simultaneous 3D reconstruction and 6D object pose estimation from a single monocular image is an inherently ill-posed problem. In industrial settings, however, multiple instances of an object are often randomly arranged in bins, implicitly providing several views of the same object within a single image. We show that this implicit multi-view geometry can be exploited to simultaneously reconstruct the object in 3D and estimate the 6D pose of each visible object instance. We present MooMIns, a new Gaussian-splatting-based approach that inverts the original Gaussian splatting formulation: instead of rendering a single scene from multiple cameras, we render multiple object instances from a single camera. Our method is initialized with SAM3 instance segmentation masks and a modified Structure from Motion (SfM) pipeline. In contrast to learned monocular depth estimation, we perform true geometry-based reconstruction from image evidence, avoiding hallucinations caused by training data priors. We evaluate MooMIns on synthetic and real bin-picking scenarios, and demonstrate accurate reconstruction of previously unseen objects as well as reliable pose estimation of individual instance

02.
PLOS Computational Biology 2026-06-10

A mean-field model of neural networks with PV and SOM interneurons reveals connectivity-based mechanisms of gamma oscillations

by Farzin Tahvili, Martin Vinck, Matteo Di Volo Classic theoretical models of cortical oscillations are based on the interactions between two populations of excitatory and inhibitory neurons. Nevertheless, experimental studies and network simulations suggest that interneuron subclasses such as parvalbumin (PV) and somatostatin (SOM) exert distinct control over oscillatory dynamics. Yet, we lack a theoretical understanding of the mechanisms underlying oscillations in E-PV-SOM circuits and of the differences with respect to the classical mechanisms for oscillations in simpler E–I networks. Here, we derive a biologically realistic mean-field model of a canonical three-population E-PV-SOM circuit. This model robustly generates oscillations whose features are consistent with experimental observations, including the relative timing of PV and SOM activity and the effects of optogenetic perturbations. By reducing the model to a linear analytical form, we demonstrate that gamma oscillations emerge directly from the cell-specific connectivity of the three-population circuit. This connectivity motif alone accounts for experimentally observed phase relationships, with PV activity consistently leading that of SOM neurons. Together, this mean field model identifies a distinct structural mechanism giving rise to oscillations in canonical E–PV–SOM circuits and provides theoretical primitives for constructing large-scale, cell-type-specific models of cortical dynamics.

03.
arXiv (CS.CV) 2026-06-11

i1: A Simple and Fully Open Recipe for Strong Text-to-Image Models

Diffusion models have consistently driven progress in text-to-image generation. However, it is challenging to attribute recent progress to specific modeling and data choices: state-of-the-art open-weight models provide limited ablations, and do not disclose their training data and full training details. The research community needs fully open (weights, data, and code) models as a foundation for further research; yet existing fully open models still fall significantly short of leading models in performance. In this project, we conduct a systematic investigation of the modeling and data design choices in text-to-image diffusion training and inference with 300+ controlled experiments totaling 700K+ TPU v6e hours. Our experiments highlight several empirical findings (e.g., equal weighting is a strong default for mixing curated datasets) and simple design decisions (e.g., larger text encoder adapters improve performance with minimal added parameters) for training strong models. Guided by these insights, we train i1, a 3B-parameter text-to-image diffusion model using only publicly available datasets. i1 is competitive with leading models on five representative benchmarks (GenEval, DPG, PRISM, CVTG-2K, and LongText), and outperforms the best existing fully open model by 29.5 absolute percentage points on average. We provide the i1 checkpoints, training and inference code, and the data processing pipeline. Together, our findings and the i1 recipe establish a practical foundation for future open research in text-to-image diffusion models. Our code is available at https://github.com/zlab-princeton/i1.

04.
arXiv (CS.LG) 2026-06-12

Multimodal Graph Negative Learning

arXiv:2606.12863v1 Announce Type: new Abstract: Multimodal attributed graphs (MAGs) integrate graph topology with heterogeneous modality attributes, such as text and images, thereby enabling richer modeling of complex relational systems. However, such expressiveness also makes learning on MAGs depend on multiple semantic sources, including structural topology, textual and visual attributes, each of which can be regarded as a branch for node representation. Node-level branch semantic imbalance arises when these branches differ across nodes in semantic informativeness and reliability: a branch that provides discriminative semantics for one node may mislead another due to bias in modality quality or structural context. Existing methods often mitigate such heterogeneity through cross-branch agreement or alignment, implicitly treating the dominant prediction as reliable supervision. When the dominant branch is biased, forced imitation may propagate its bias to other branches and suppress original semantics that are useful for classification. We propose GraphMNL, a graph-aware multimodal negative learning framework that addresses this issue by using Negative Learning as cross-branch guidance. Instead of forcing inferior branches to imitate a teacher prediction, the model teaches them which classes a node is unlikely to belong to. GraphMNL builds a branch library, identifies dominant and inferior branches via graph-aware reliability arbitration, gates unstable transfer, and applies target-preserving negative learning over non-target classes. This design decouples target supervision from branch guidance so that supervised losses learn the correct class, while Negative Learning suppresses unlikely alternatives when branch agreement is unreliable. Through the comprehensive experimental evaluation, GraphMNL achieves the best performance on Grocery datasets with 72.47% accuracy and 76.60 F1 score on Reddit M datasets.

05.
arXiv (math.PR) 2026-06-16

Risk-averse mean field games: exploitability and non-asymptotic analysis

arXiv:2301.06930v5 Announce Type: replace-cross Abstract: In this paper, we use mean field games (MFGs) to investigate approximations of $N$-player games ($N$pGs) with uniformly symmetrically continuous heterogeneous closed-loop actions. To incorporate agents' risk aversion (beyond the classical expected utility of total costs), we use an abstract evaluation functional for their performance criteria. Centered around the notion of exploitability, we conduct non-asymptotic analysis on the approximation capability of MFGs from the perspective of state-action distributions without requiring the uniqueness of equilibria. Under suitable assumptions, we first show that scenarios in the $N$pGs with large $N$ and small average exploitabilities can be well approximated by approximate solutions of MFGs with relatively small exploitabilities. We then show that $\delta$-mean field equilibria can be used to construct $\varepsilon$-equilibria in $N$pGs. Furthermore, in this general setting, we prove the existence of mean field equilibria. This proof reveals a possible avenue for incorporating penalization for randomized action into MFGs.

06.
arXiv (CS.CL) 2026-06-11

M4FC: a Multimodal, Multilingual, Multicultural, Multitask Real-World Fact-Checking Dataset

Existing real-world datasets for multimodal fact-checking have multiple limitations: they contain few instances, cover on only one or two languages, focus only on one task, or rely on external news article sets for sourcing true claims. To address these shortcomings, we introduce M4FC, a new real-world dataset comprising 4,982 images paired with 6,980 claims. The images, verified by professional fact-checkers from 22 organizations, represent a diverse range of cultural and geographic contexts. Each claim is available in one or two out of ten languages. M4FC spans six multimodal fact-checking tasks: visual claim extraction, claimant intent prediction, fake image detection, image contextualization, location verification, and verdict prediction. We provide baseline results for all tasks and analyze how combining intermediate tasks affects verdict prediction performance. We make our dataset and code publicly available.

07.
arXiv (CS.LG) 2026-06-18

A Cross-Model VLM-Judge Protocol for Single-Image 3D Mesh Quality (and Why Cheap Proxies Fall Short)

arXiv:2606.18451v1 Announce Type: new Abstract: Single-image-to-3D generators are improving quickly, but there is no agreed, human-free way to tell whether one generated mesh is better than another. Practitioners commonly rely on cheap automatic proxies (render-space CLIP similarity and mesh geometry-validity statistics), yet how well these track perceived quality is unestablished. We make two contributions. First, we propose and validate a reproducible VLM-judge evaluation protocol: a fixed 24-view headless render rig, two independent vision-language judge families, and a mandatory position-bias correction that queries both presentation orders and keeps only order-consistent verdicts. The two judge families agree substantially with each other (Cohen's kappa = 0.66), well above the chance-agreement floor. Second, using this protocol as the reference, we show the cheap proxies do not substitute for it. Geometry validity is only a weak signal on average (because, as we show, it is bimodal) and stays below our pre-registered target, while render-CLIP is at chance. A learned Bradley-Terry head collapses onto a single manifoldness statistic (giving render-CLIP a negative weight) and matches geometry-only exactly, so learning the feature weights buys nothing. The proxy is also bimodal: it is significantly above chance on contrasts with visible geometric defects but at chance on ambiguous contrasts, consistent with geometry validity tracking the judge only when the defect is visually salient. We therefore recommend the VLM-judge protocol as a reliable, reproducible evaluator under the conditions tested (two feed-forward generators on Google Scanned Objects, with a face-drop degradation regime) and advise against geometry/CLIP proxies as optimization targets.

08.
arXiv (CS.LG) 2026-06-16

Stochastic Schrödinger Diffusion Models for Pure-State Ensemble Generation

arXiv:2605.03573v3 Announce Type: replace-cross Abstract: Quantum machine learning increasingly relies on pure-state representations, motivating generative models that sample directly in quantum representation space rather than perturbing classical inputs and re-encoding. We introduce Stochastic Schrödinger Diffusion Models (SSDMs), a score-based generative framework that defines diffusion, scores, and reverse-time sampling intrinsically on the complex projective manifold $\mathbb{CP}^{d-1}$ under the Fubini–Study metric. SSDMs combine a Riemannian Ornstein–Uhlenbeck forward diffusion with a stochastic Schrödinger realization, and learn reverse-time dynamics driven by the Riemannian score. Our central technical contribution is a local-time learning objective that exploits the local Euclidean OU limit of intrinsic manifold diffusions in Fubini-Study normal coordinates to obtain an analytic teacher score, bypassing the intractable transition densities that limit existing Riemannian score-based models. Across synthetic, physics-inspired (TFIM, XXZ), and quantum feature-state benchmarks up to $14$ qubits, SSDMs match target pure-state ensembles by orders of magnitude on MMD and observable statistics over both ambient Euclidean and matched Riemannian score-based baselines, and improve representation-level diagnostics for downstream quantum kernel methods.

09.
arXiv (CS.AI) 2026-06-11

DecompSR: A dataset for decomposed analyses of compositional multihop spatial reasoning

arXiv:2511.02627v3 Announce Type: replace Abstract: We introduce DecompSR, decomposed spatial reasoning, a large benchmark dataset (over 5m datapoints) and generation framework designed to analyse compositional spatial reasoning ability. The generation of DecompSR allows users to independently vary several aspects of compositionality, namely: productivity (reasoning depth), substitutivity (entity and linguistic variability), overgeneralisation (input order, distractors) and systematicity (novel linguistic elements). DecompSR is built procedurally in a manner which makes it is correct by construction, which is independently verified using a symbolic solver to guarantee the correctness of the dataset. DecompSR is comprehensively benchmarked across a host of Large Language Models (LLMs) where we show that LLMs struggle with productive and systematic generalisation in spatial reasoning tasks whereas they are more robust to linguistic variation. DecompSR provides a provably correct and rigorous benchmarking dataset with a novel ability to independently vary the degrees of several key aspects of compositionality, allowing for robust and fine-grained probing of the compositional reasoning abilities of LLMs.

10.
arXiv (CS.LG) 2026-06-18

Fair Online Resource Allocation

arXiv:2606.18679v1 Announce Type: cross Abstract: We study the problem of fair online resource allocation, motivated by applications such as refugee resettlement and airline scheduling, where agents arrive sequentially and must be assigned to facilities with limited capacities. We introduce a model that maximizes the overall welfare subject to resource constraints and a Lipschitz fairness requirement, which ensures that similar agents arriving in the same batch receive similar expected outcomes. We first analyze the offline problem, proving that the value of the optimal fair allocation is at least an $\Omega(1/\gamma)$ fraction of the optimal unfair allocation, where $\gamma$ is the fairness coefficient, thereby bounding the price of fairness. For the online setting, we propose an algorithm based on dual mirror descent that enforces fairness constraints within batches while estimating optimal dual variables. We prove that this algorithm achieves sublinear regret relative to the optimal offline fluid benchmark. Finally, we validate our theoretical results using real-world data from the Refugee Economies Programme, demonstrating the algorithm's performance and examining the trade-offs between welfare maximization and fairness enforcement.

11.
arXiv (CS.LG) 2026-06-17

Multi-Source Cybersecurity Logs: An ATT&CK-Labeled Dataset and SLM Evaluation

arXiv:2606.18190v1 Announce Type: cross Abstract: Multi-stage cyberattacks span system, network, and browser logs. Detecting them requires correlating events across all three sources. Machine learning methods can learn these cross-source patterns, but they need labeled multi-source data. Existing public datasets fall short. Network-only datasets such as CICIDS and UNSW-NB15 miss host and browser activity. Host-focused datasets such as LMDG and CICAPT-IIoT lack browser telemetry. ATLAS includes all three sources but labels events only as malicious or benign, without MITRE Adversarial Tactics, Techniques, and Common Knowledge (ATT&CK) technique granularity. No public dataset combines all three sources with per-entry ATT&CK technique labels. We close the gap by building a multi-source log dataset of 870 sessions (70 attack, 800 benign) and approximately 2.3 million events. We captured system, network, and browser activity simultaneously on Windows endpoints. We labeled malicious events with ATT&CK technique IDs, covering 12 tactics and 53 techniques. We generated all attack data using real tools, including Remote Access Trojan (RAT), Command and Control (C2) tunnels, and cloud exfiltration. To demonstrate learnability, we fine-tuned three Small Language Models (SLMs) (Qwen2.5-1.5B, Llama-3.2-3B, Phi-4-Mini) using Low-Rank Adaptation (LoRA). We compared each against its base variant across ten metrics on two tasks: chunk classification and ATT&CK technique identification. Fine-tuning improved every model on every metric. Chunk classification accuracy rose from approximately 8% in the base variants to between 90% and 97% after fine-tuning. Technique identification remained challenging, with the best exact-match accuracy at 42%, although high partial-match scores show the models captured most of the underlying reasoning.

12.
arXiv (CS.CL) 2026-06-16

From Awareness to Adherence: Bridging the Context Gap in Spoken Dialogue Systems via Context-Aware Decoding

Despite the success of end-to-end (E2E) spoken dialogue systems, maintaining strict context adherence in multi-round conversations remains a challenge. While prior works attribute these failures to models forgetting dialogue history, we highlight an equally critical but overlooked bottleneck: a gap between latent context awareness and active adherence. Although models internally recognize relevant past utterances, strong parametric priors often overshadow these signals during decoding. To bridge this gap, we propose an audio-adapted Context-Aware Decoding (CAD) approach. By leveraging internal attention mechanisms to isolate key historical rounds, our approach contrasts output distributions with and without this key context during inference, directly amplifying multimodal contextual signals. Evaluations on the Audio MultiChallenge benchmark demonstrate significant improvements in Semantic Memory and Self Coherence subtasks, successfully enforcing strict, context-faithful adherence.

13.
arXiv (CS.AI) 2026-06-16

Phantoms and Disclosures: a Causal Framework for Auditing Synthetic Data

arXiv:2606.16952v1 Announce Type: cross Abstract: The rapid adoption of generative AI and Large Language Models (LLMs) has spurred interest in synthetic data as a privacy-preserving alternative to sensitive real-world datasets. However, generating high-utility synthetic data often carries the risk of memorizing and regurgitating private information from the training corpus. In this work, we present a customizable empirical auditing framework designed to detect and explain such data disclosures. Our framework introduces a mechanism to distinguish between "true disclosures"-where the system directly reproduces a user's information-and "phantom disclosures''-where the system incidentally generates a user's data. By partitioning input data into training and holdout sets and applying rigorous statistical hypothesis testing, we determine if observed disclosures are consistent with strict privacy baselines, such as zero-learning or specific Differential Privacy (DP) bounds. Crucially, this approach requires no model access, no canary insertion, and no reference model training -only the synthetic output and a held-out control set. We demonstrate that this framework effectively functions as a membership inference attack, providing empirical lower bounds on privacy leakage that are tighter than prior data-based auditing methods. Our approach is model-agnostic, applies to any synthetic data generation mechanism, and requires orders of magnitude fewer computational resources than shadow-model or canary-based alternatives.

14.
arXiv (quant-ph) 2026-06-16

Worst-case depth hierarchy for shallow quantum circuits

arXiv:2606.16425v1 Announce Type: new Abstract: Circuit depth is a central resource in complexity theory. While bounded-depth classical circuits admit well-understood hierarchy theorems, the internal structure of constant-depth quantum computation remains comparatively unexplored. We prove an explicit depth hierarchy theorem for $\mathsf{QNC}^0$. For each $d\ge 12$, we construct a family of two-round interactive problems on which no depth-$(d-1)$ quantum circuit can achieve near-perfect success, regardless of gate set, circuit size, or ancillary qubits. In contrast, we prove that our construction admits realizations by simple bounded fan-in quantum circuits of depth larger than $d$ by a small constant factor. Moreover, all bounded fan-in classical circuits of sublogarithmic depth (in the input size) fail to achieve perfect success on these tasks for every $d$, yielding a hierarchy of problems that show unconditional quantum advantage of $\mathsf{QNC}^0$ over $\mathsf{NC}^0$. A key obstacle is the scarcity of lower bound techniques for quantum circuits. To address this, we develop methods to analyze how depth affects a circuit's ability to realize nonlocal correlations amongst its output qubits in a fine-grained manner. Our approach exploits the correspondence between constraint systems and nonlocal games, translating group-theoretic constructions into rigid operator-valued constraint systems and then into non-local games. In particular, we construct constraint systems whose unique faithful operator-valued solutions require every perfect strategy, and every near-perfect strategy to a fixed precision, to implement multi-controlled phase operations. This reduces to a nonlocal unitary-synthesis problem, yielding depth lower bounds for both shallow quantum and classical circuits. These results show that increasing depth strictly increases computational power within $\mathsf{QNC}^0$, establishing a genuinely quantum hierarchy.

15.
arXiv (CS.LG) 2026-06-17

Recursive Scaling in Masked Diffusion Models

arXiv:2606.18022v1 Announce Type: new Abstract: Masked diffusion models (MDMs) have recently emerged as a promising paradigm for sequence generation. Scaling MDMs is conventionally achieved by increasing the parameter count or the number of denoising steps. We introduce Recursive Masked Diffusion Models (R-MDMs), which add recursive depth as a third scaling axis by repeatedly applying the same denoising transformer within each diffusion step. Recursion enables iterative refinement of the output through parameter reuse, increasing effective model depth without increasing parameter count. Across structured generation tasks, including Sudoku and Countdown, we show that R-MDMs achieve substantially improved parameter efficiency: a model with $L$ recursive iterations often matches the performance of non-recursive baselines with roughly $L\times$ more parameters. Moreover, recursive refinement can partially substitute for additional denoising steps, allowing recursive models to reach the same generation quality with fewer forward passes at inference time. These results suggest that recursive depth is a practically useful scaling mechanism for MDMs, improving both parameter efficiency and the allocation of test-time compute.

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

Unifying Post-hoc Explanations of Knowledge Graph Completions

arXiv:2507.22951v2 Announce Type: replace Abstract: Knowledge Graphs organize information as entity-relation-entity triples, enabling machine learning models to predict plausible missing triples in a task known as Knowledge Graph Completion (KGC). Post-hoc explainability for KGC addresses the problem of identifying which triples most influence the predictions of machine learning models. Currently, the field lacks formalization and consistent evaluations, hindering reproducibility and cross-study comparisons. This paper argues for a unified taxonomy for post-hoc explainability in KGC. First, we propose a characterization of post-hoc explanations via multi-objective optimization that unifies existing post-hoc explainability algorithms in KGC and the explanations they produce, balancing explanation effectiveness and conciseness. Next, we examine improved evaluation protocols based on popular metrics, such as Mean Reciprocal Rank and Hits@k, through illustrative experiments. Finally, we stress the importance of interpretability as the ability of explanations to address queries meaningful to end users. By unifying methods and discussing evaluation standards, this work puts forward a case for more reproducible and impactful research in KGC explainability.

18.
arXiv (CS.CV) 2026-06-16

G2IA: Geometry-Guided Instance-Aware Retrieval and Refinement for Cross-Modal Place Recognition

Cross-modal place recognition (CMPR) enables camera-only robots to localize against pre-built LiDAR maps in autonomous navigation scenarios. This image-to-point-cloud setting is challenged by two coupled ambiguities: the modality gap between perspective RGB appearance and sparse metric geometry, and perceptual aliasing among urban places with similar roads, facades, intersections, and object arrangements. Instead of treating CMPR as a single global descriptor matching problem, we argue that reliable retrieval requires both geometry-aware representation alignment and fine-grained candidate verification. In this paper, we propose G2IA, a geometry-guided instance-aware framework for image-to-point-cloud place recognition. In the retrieval stage, visual geometry priors from VGGT and instance features are integrated to construct place descriptors that are more compatible with LiDAR-derived map representations. In the refinement stage, the retrieved candidates are re-ranked by explicitly verifying whether local instance shapes and their relative spatial layouts are consistent across modalities. Experiments on public benchmarks demonstrate that G2IA consistently improves image-to-point-cloud place recognition under different localization thresholds, and exhibits strong cross-dataset generalization.

19.
arXiv (CS.LG) 2026-06-19

Variational Consensus Monte Carlo for Bayesian Mixture

arXiv:2606.19643v1 Announce Type: cross Abstract: Motivated by the privacy, sensitivity and sharing limitations of health data, we present a comprehensive pipeline for inference of Bayesian mixture models within a federated learning setting, i.e. when data cannot be fully shared or pooled across compute nodes. We adopt a Consensus Monte Carlo (CMC) approach, in which an MCMC algorithm is run independently within each data silo to estimate local posterior distributions, which are then aggregated to approximate the posterior over the full data. The variational CMC approach of Rabinovich, Angelino and Jordan (2015) [1] frames the aggregation step as a variational inference problem, but their application to mixtures assumes the number of clusters and key mixture parameters to be known. Our main methodological contributions are: (i) an extension of variational CMC to over-fitted Bayesian mixture models that infer the number of clusters and all model parameters, without requiring conjugacy; (ii) novel cluster-matching algorithms suitable for cross-silo settings in which not every cluster appears in each local dataset; (iii) a number of inference strategies for the aggregation step, matched to different federated learning constraints; and (iv) guidelines for choosing among these in practice. A comprehensive simulation study validates the framework and allows us to compare to state-of-the-art federated learning alternatives. Notably, we show that when the composition of local datasets reflects the underlying clustering structure in the data, our approach can recover small clusters with greater accuracy than standard MCMC applied to the pooled data. We illustrate the framework on large-scale electronic health record data, identifying multi-morbidity patterns in a British geriatric population.

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

Learning to Prompt: Improving Student Engagement with Adaptive LLM-based High-School Tutoring

LLMs can personalize education, although current static-prompt tutoring systems struggle to adapt to diverse academic disciplines. We develop and test a system with subject-aware prompting, based on 14 pedagogical features (e.g., tutor scaffolding, student understanding) extracted from raw transcripts. We first train a prompt routing model in a simulation environment, and then deploy it for online adaptation with actual high-school students. The simulation benchmark shows the router outperforming two static baselines ($0.694$ vs. $0.647$ and $0.64$, $p

21.
arXiv (CS.LG) 2026-06-11

On Subquadratic Architectures: From Applications to Principles

arXiv:2606.12364v1 Announce Type: new Abstract: Transformers dominate modern sequence modeling, but their quadratic attention incurs substantial computational cost. Subquadratic architectures offer a scalable alternative. However, it remains unclear which designs yield the most effective sequence models. We compare three leading approaches: xLSTM, Mamba-2, and Gated DeltaNet. We evaluate these models on tasks with complex dependencies: (1) code-model pre-training, (2) distillation of code models from large language models, and (3) pre-training of time-series foundation models. Across these settings, xLSTM delivers the strongest overall performance. To explain xLSTM's advantage, we present a unified formulation and analyze the underlying architectural mechanisms, focusing on state tracking and memory dynamics. Our results show that xLSTM enables more flexible and stable memory correction via its gating scheme. We corroborate these findings on controlled synthetic length-generalization tasks. Overall, our findings indicate that xLSTM's gains on complex tasks stem from robust state tracking and accumulation.

22.
arXiv (CS.CV) 2026-06-17

Pulling The REINS: Training-Free Safety Alignment of Video Diffusion Models via Representation Steering

Open-weight video diffusion models can generate photorealistic unsafe content, from violence to misinformation, yet existing defenses either require expensive safety fine-tuning that degrades general capability, or apply external filters that are trivially bypassed by adversarial prompts. We present REINS (REpresentation-space INference-time Safety steering), a training-free method that aligns video diffusion models at inference time by steering their internal representations toward safe generation. Our key finding is that safety-relevant structure is linearly encoded in the hidden-state activations of video diffusion transformers, and a single direction, discovered via Supervised PCA on binary safety labels, suffices to separate safe from unsafe generation trajectories. At inference, adding this direction to hidden states at an intermediate transformer layer redirects generation from harmful content to semantically related safe alternatives, with no weight updates, no concept enumeration, and negligible computational overhead. Through mechanistic analysis, we reveal that while safety information accumulates monotonically with transformer depth, steering effectiveness peaks at intermediate layers (~50% depth), exposing a fundamental tradeoff between information availability and downstream propagation capacity. We evaluate REINS across 9 video diffusion models, multiple parameter scales (1.3B-5B), and both text-to-video and image-to-video generation, to our knowledge, the broadest safety evaluation suite in the video generation literature.

23.
arXiv (CS.AI) 2026-06-17

AnchorKV: Safety-Aware KV Cache Compression via Soft Penalty with a Refusal Anchor

arXiv:2606.17872v1 Announce Type: cross Abstract: Large language models (LLMs) outperform earlier architectures on generative inference and long-context tasks, but their large size introduces significant challenges in memory usage, energy cost, and on-device deployment. Since scaling pre-trained language models improves downstream capability [zhao2023survey], the key-value (KV) cache becomes a dominant inference bottleneck. Recent KV cache compression methods [jo2025fastkv,li2024snapkv,zhou2024dynamickv] reduce this cost by retaining only a subset of attention-relevant tokens. However, while these approaches preserve accuracy on benign workloads, their compression policies either fail to defend against jailbreak attacks [jiang2024robustkv] or degrade safety alignment under aggressive eviction. We propose AnchorKV, a drop-in modification to KV cache compression that biases token retention scores away from directions in key space associated with harmful prompts. AnchorKV constructs an offline safety anchor by adapting a difference-of-means representation engineering approach [arditi2024refusal,zou2023representation] to the layer-specific key projection space used in KV caching. Based on this anchor, a soft penalty token selection rule trades a small amount of utility for substantially improved safety alignment, while reducing to the original compressor when the penalty is zero.

24.
arXiv (CS.AI) 2026-06-19

StreamKL: Fast and Memory-Efficient KL Divergence for Boosting Attention Distillation

arXiv:2606.20005v1 Announce Type: cross Abstract: Attention distillation, which trains one attention distribution to match another by minimizing their Kullback-Leibler (KL) divergence, is widely used in knowledge distillation, model compression, continual learning, and sparse-attention LLM training. However, existing approaches materialize both attention distributions before computing the KL reduction, incurring $O(N_QN_K)$ memory and IO costs that become prohibitive at long context lengths. We present StreamKL, the first fused GPU primitive for attention KL divergence that eliminates this quadratic materialization. StreamKL derives a novel online formulation for the coupled two-distribution KL reduction, enabling a single one-pass forward kernel that streams query-key tiles through on-chip SRAM. For the backward pass, StreamKL recomputes attention probabilities tile-by-tile, avoiding storage of quadratic intermediates. We further design and implement efficient GPU kernels with dedicated optimizations. Experiments show StreamKL delivers up to $43\times$ and $14\times$ speedups over baseline methods in the forward and backward passes, respectively. Most importantly, StreamKL reduces the extra HBM footprint of attention distillation from $O(N_QN_K)$ to $O(1)$, enabling long-context distillation on a single GPU.

25.
arXiv (CS.CL) 2026-06-12

EvoArena: Tracking Memory Evolution for Robust LLM Agents in Dynamic Environments

Large language model (LLM) agents have achieved strong performance on a wide range of benchmarks, yet most evaluations assume static environments. In contrast, real-world deployment is inherently dynamic, requiring agents to continually align their knowledge, skills, and behavior with changing environments and updated task conditions. To address this gap, we introduce EvoArena, a benchmark suite that models environment changes as sequences of progressive updates across terminal, software, and social domains. We further propose EvoMem, a patch-based memory paradigm that records memory evolution as structured update histories, enabling agents to reason about environmental evolution through changes in their memory. Experiments show that current agents struggle on EvoArena, achieving an average accuracy of 39.6% across evolving terminal, software, and social-preference domains. EvoMem consistently improves performance, yielding an average gain of 1.5% on EvoArena and also improving standard benchmarks such as GAIA and LoCoMo by 6.1% and 4.8%. Beyond individual tasks, EvoMem further improves chain-level accuracy by 3.7% on EvoArena, where success requires completing a consecutive sequence of related evolutionary subtasks. Mechanistic analysis shows that EvoMem improves evidence capture in the memory, indicating better preservation of complete evolving environment states. Our results highlight the importance of modeling evolution in both evaluation and memory for reliable agent deployment.