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.
medRxiv (Medicine) 2026-06-11

A continental-scale scenario modelling framework for evaluating infant RSV immunisation strategies across Europe

Background. The recent approval of long-acting monoclonal antibodies (la-mAbs) and a maternal vaccine (MV) in the EU enables universal RSV prevention in infants. Modelling studies are widely used to quantify the population-level impact of alternative immunisation strategies. However, existing assessments of new RSV immunisation products focus on national or sub-national settings. Methods. We developed an age-stratified, stochastic compartmental model of RSV transmission for 28 EU/EEA countries. It combines literature-based parameters on RSV natural history and product efficacy with country-specific demographic and contact patterns. After model calibration against age- and country-specific RSV hospitalisation rates, we designed scenarios for both la-mAbs and MV at four coverage levels, with and without catch-up immunisation for infants under six months at season onset. We then evaluated each scenario against a no-immunisation baseline. Results. At 95% coverage, the cross-country median reduction in RSV hospitalisations over one season in infants under 12 months is 29.9% for la-mAbs (country median range: 27.7-33.9%) and 22.4% for MV (20.0-25.6%), scaling linearly with coverage. Out of all averted hospitalisations, 78.3% (90% CI: [67.3, 92.7]%) are concentrated in infants aged 0-2 months for la-mAbs and 72.7% (90% CI: [61.4, 88.6]%) for MV. A catch-up campaign nearly doubles the overall reduction in RSV hospitalisations. Conclusions. Despite country-specific heterogeneities, impact of la-mAbs and MV is comparable across settings and herd-immunity effects are largely negligible. This supports harmonised European guidelines on coverage targets. Seasonal catch-up campaigns emerge as an effective lever to maximise the impact of immunisation programmes.

02.
Nature (Science) 2026-06-16

Mathematicians are developing rules for AI use — other fields should follow

作者: 未知作者

The mathematics community is right to call for transparency, integrity and fairness to be protected when AI tools are used. Researchers in other disciplines could learn from this approach. The mathematics community is right to call for transparency, integrity and fairness to be protected when AI tools are used. Researchers in other disciplines could learn from this approach.

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

Observable Patterns Are Not Explanations: A Causal-Geometric Analysis of Latent Reasoning Models

Latent reasoning models (LRMs) replace explicit chain-of-thought with continuous thoughts. Recent work treats observable latent-state patterns, such as BFS-like frontiers and decodable arithmetic computation, as evidence for internal reasoning mechanisms. Evaluating two LRMs (Coconut and CODI) against controls lacking the proposed recurrence or curriculum, we find these patterns also appear in the controls and do not always causally affect behavior. Causal interventions reveal that latent-thought utilization is not binary but graded, scaling with a thought's causal effect on model behavior. Geometric analyses reveal this effect concentrates in low-rank directions whose step-to-step geometry grows more structured as their behavioral influence increases. Latent thoughts should therefore be treated as hidden computation, not hidden explanation: decodability, attention, or static structure alone cannot establish mechanism. LRM interpretability thus requires matched controls and causal tests.

05.
arXiv (CS.CL) 2026-06-12

Self-Distillation Zero: Self-Revision Turns Binary Rewards into Dense Supervision

Current post-training methods in verifiable settings fall into two categories. Reinforcement learning (RLVR) relies on binary rewards, which are broadly applicable and powerful, but provide only sparse supervision during training. Distillation provides dense token-level supervision, typically obtained from an external teacher or using high-quality demonstrations. Collecting such supervision can be costly or unavailable. We propose Self-Distillation Zero (SD-Zero), a method that is substantially more training sample-efficient than RL and does not require an external teacher or high-quality demonstrations. SD-Zero trains a single model to play two roles: a Generator, which produces an initial response, and a Reviser, which conditions on that response and its binary reward to produce an improved response. We then perform on-policy self-distillation to distill the reviser into the generator, using the reviser's token distributions conditioned on the generator's response and its reward as supervision. In effect, SD-Zero trains the model to transform binary rewards into dense token-level self-supervision. On math and code reasoning benchmarks with Qwen3-4B-Instruct and Olmo-3-7B-Instruct, SD-Zero improves performance by at least 10% over the base models and outperforms strong baselines, including Rejection Fine-Tuning (RFT), GRPO, and Self-Distillation Fine-Tuning (SDFT), under the same question set and training sample budget. Extensive ablation studies show two novel characteristics of our proposed algorithm: (a) token-level self-localization, where the reviser can identify the key tokens that need to be revised in the generator's response based on reward, and (b) iterative self-evolution, where the improving ability to revise answers can be distilled back into generation performance with regular teacher synchronization. Code: https://github.com/princeton-pli/Self-Distillation-Zero.

06.
arXiv (CS.CL) 2026-06-18

Notation Matters: A Benchmark Study of Token-Optimized Formats in Agentic AI Systems

Large language models in Agentic AI systems consume tool schemas and execution results and emit tool invocations as structured data. The default language for that exchange, JSON, was designed for application-to-application interchange rather than token efficiency, so its structural elements impose substantial token overhead. Recent work proposes token-optimized alternatives such as TOON (Token-Oriented Object Notation) and TRON (Token Reduced Object Notation) as more compact replacements, but these formats have been evaluated only on isolated comprehension or generation tasks. Whether their token reductions hold inside end-to-end agentic loops therefore remains an open question. We evaluate TOON and TRON on four agentic benchmarks (BFCL, MCPToolBenchPP, MCP-Universe, StableToolBench) and five open-weight LLMs, decoupling input compression from output compression to measure comprehension and generation independently. TRON reduces tokens by up to 27% with accuracy within 14pp of the JSON baseline. TOON achieves up to 18% reduction at a similar 9pp accuracy cost, but additionally cascades on multi-turn parsing failures and collapses parallel tool-call output for most models. The code is available at: https://github.com/lkutschka/notation-matters

07.
arXiv (CS.LG) 2026-06-19

Topological Data Analysis for High-Dimensional Dynamic Process Monitoring

arXiv:2606.20443v1 Announce Type: cross Abstract: Real-time process monitoring requires methods that extract actionable information from high-dimensional time-series data. In this work, we present a new approach for process monitoring that combines tools of topological data analysis (TDA) and machine learning. In the proposed approach, we represent multivariate time-series data as manifolds and use topological descriptors to summarize the structure of such data; we then use a neural ordinary differential equation to learn the dynamic evolution of the topological structure of the system. Using real data from an industrial process, we show that this trajectory-based event detection approach is effective at detecting diverse types of events. We contrast this approach against reconstruction-based approaches such as principal component analysis and autoencoders and against a trajectory-based approach that uses Koopman autoencoders.

08.
arXiv (CS.AI) 2026-06-18

As You Wish: Mission Planning with Formal Verification using LLMs in Precision Agriculture

arXiv:2606.18519v1 Announce Type: cross Abstract: Though robotic systems are now being commercialized and deployed in various industries, many of these systems are highly specialized and often require an advanced skill set to operate and ensure they perform as instructed. To mitigate this problem, we recently introduced a mission planner leveraging LLMs to synthesize mission plans in precision agriculture based on mission descriptions provided in natural language. While the system demonstrates impressive performance, it also suffers from the inherent ambiguities of natural language. In this paper, we extend our system to address this issue by introducing multiple feedback loops in the planning architecture that leverage linear temporal logic (LTL) to ensure the mission planning system meets the specifications formulated by the user while still using natural language. To mitigate potential bias, this is achieved by using two different commercial LLMs in charge of the specification and verification subtasks. Through extensive experiments, we highlight the strengths and limitations of integrating mission verification into a fully autonomous pipeline, particularly regarding an LLM's ability to generate valuable LTL formulas, and show how our proposed implementation addresses and solves these challenges.

09.
arXiv (CS.AI) 2026-06-16

Let Them Steal: Trapping Large Language Model Extraction Attacks with Knowledge Honeypot

arXiv:2606.15810v1 Announce Type: cross Abstract: Large language models deployed as commercial APIs are vulnerable to model extraction attacks, while existing defenses either act too late or degrade utility for legitimate users. We propose Knowledge Trap, a defense that redirects extraction attacks toward low-transferability knowledge through a Honeypot Knowledge Graph (HKG) and breadcrumb-guided exploration. Instead of blocking queries or perturbing outputs, Knowledge Trap consumes the attacker's limited query budget on knowledge with negligible downstream utility while preserving benign-user performance. Experiments in medical and financial domains show that Knowledge Trap reduces surrogate Agreement by 6.2\% on average without degrading legitimate-user accuracy, outperforming existing defenses that impose measurable user impact. These results suggest that defending knowledge-space traversal is a practical direction for mitigating LLM extraction attacks.

10.
arXiv (CS.LG) 2026-06-16

Bayesian Networks with Latent Time Embedding for Stage-Aware Causal Modeling of Alzheimer's Disease Progression

arXiv:2606.15784v1 Announce Type: new Abstract: Alzheimer's disease (AD) progression is often described through the amyloid-tau-neurodegeneration, or AT(N), cascade. However, most longitudinal models represent this cascade either as a fixed sequence of biomarkers or as a black-box forecasting task. This makes it difficult to determine when biologically guided biomarker relationships influence future regional pathology. In this study, we introduce Bayesian Networks with Latent Time Embedding (BN-LTE), a Bayesian structural framework for stage-aware modeling of AD progression. BN-LTE estimates disease pseudotime from baseline biomarker profiles and constrains directed dependencies according to biologically plausible AT(N) ordering. Posterior spline-varying structural equations are then used to link initial multimodal measurements with future annualized regional tau-PET change. Across repeated subject-disjoint evaluations using ADNI data, BN-LTE shows strong spatial reconstruction of tau progression compared with the included forecasting baselines. Beyond spatial reconstruction, BN-LTE recovers posterior stage-varying AT(N)-constrained effects and identifies a mid-pseudotime window of amyloid sensitivity. This window is supported by model-implied g-formula contrasts, root-adjusted AIPW, mechanism-sensitive ablations, and robustness analyses across spline and prior specifications. Overall, these findings position BN-LTE as a Bayesian structural framework for forecasting tau progression while examining stage-dependent AT(N)-cascade mechanisms in observational longitudinal neuroimaging data. Our code is available at https://github.com/danleneurocom/BN-LTE.

11.
arXiv (CS.CV) 2026-06-16

Rel-Zero: Harnessing Patch-Pair Invariance for Robust Zero-Watermarking Against AI Editing

Recent advancements in diffusion-based image editing pose a significant threat to the authenticity of digital visual content. Traditional embedding-based watermarking methods often introduce perceptible perturbations to maintain robustness, inevitably compromising visual fidelity. Meanwhile, existing zero-watermarking approaches, typically relying on global image features, struggle to withstand sophisticated manipulations. In this work, we uncover a key observation: while individual image patches undergo substantial alterations during AI-based editing, the relational distance between patch pairs remains relatively invariant. Leveraging this property, we propose Relational Zero-Watermarking (Rel-Zero), a novel framework that requires no modification to the original image but derives a unique zero-watermark from these editing-invariant patch relations. By grounding the watermark in intrinsic structural consistency rather than absolute appearance, Rel-Zero provides a non-invasive yet resilient mechanism for content authentication. Extensive experiments demonstrate that Rel-Zero achieves substantially improved robustness across diverse editing models and manipulations compared to prior zero-watermarking approaches.

12.
arXiv (math.PR) 2026-06-19

Power-law hypothesis and (un)fairness of PageRank on undirected multi-type PAMs

arXiv:2606.19583v1 Announce Type: new Abstract: The preferential attachment model (PAM) describes the sequential growth of a network based on the "rich-get-richer" principle. Several versions of it have become established for modeling, e.g., citation networks, capturing a power-law degree distribution. Directed versions of the preferential attachment model where the edges are directed from the new to the old vertices have been the subject of extensive research. They have been shown to exhibit remarkable properties such as heavier tails for the limiting graph-normalized PageRank than for the in-degrees. By contrast, for the undirected version, we recently showed that PageRank has similar tails as the degree. In the present paper, we discuss the PageRank asymptotics for a multi-type version of the undirected PAM (here vertices have different colors), complementing previous results of Antunes, Bhamidi, Banerjee and Pipiras on the asymptotics of PageRank on similar directed multi-type or colored PAMs. Our studies are motivated by the aim to go beyond the rigid rule of edge orientation in directed preferential attachment models. As the main result, for the case of a finite set of colors, we show that the power-law hypothesis for PageRank is fulfilled also for the colored undirected PAM, where, by contrast to the directed case, the power-law exponent is color-dependent for some choices of the initial color distribution and the attractiveness function. For the specific case of a two-type model, we discuss implications of our results on fairness in sampling underrepresented nodes from the network.

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

Fusion Learning from Dynamic Functional Connectivity: Combining the Amplitude and Phase of fMRI Signals to Identify Brain Disorders

arXiv:2603.24603v2 Announce Type: replace-cross Abstract: Dynamic functional connectivity (dFC) derived from resting-state functional magnetic resonance imaging (fMRI) has been extensively utilized in brain science research. The sliding window correlation (SWC) method is a widely used approach for constructing dFC by computing correlation coefficients between amplitude time series of signals from pairs of brain regions. In this study, we propose an integrated approach that incorporates both amplitude and phase information of fMRI signals to improve the detection of brain disorders. Specifically, we introduce a multi-scale fusion learning framework, namely MSFL, which leverages two complementary dFC features derived from SWC and phase synchronization (PS). Here, SWC captures amplitude correlations, while PS measures phase coherence within dFC. We evaluated the efficacy of MSFL in classifying autism spectrum disorder and major depressive disorder using two publicly available datasets: ABIDE I and REST-meta-MDD, respectively. The results indicate that MSFL significantly outperforms existing comparative models. Moreover, we performed model explanation analysis using the SHAP framework, which showed that both types of dFC features from SWC and PS contribute to detecting brain disorders.

14.
arXiv (CS.LG) 2026-06-16

Cross-Silo De-Anonymization Under Local Differential Privacy: Threat Model, Phase Transition, and Coordination Necessity

arXiv:2606.16763v1 Announce Type: cross Abstract: When a person's records appear in k independent data silos, each protected by (epsilon, delta)-differential privacy, standard composition yields a valid (k*epsilon, k*delta)-DP guarantee for the joint output. This worst-case bound, however, does not answer the concrete inference question: at what k can an adversary actually identify a target person? This paper develops the information-theoretic framework needed to answer that question. We introduce cross-silo person-level DP (XSP-DP), a Pufferfish-style privacy notion whose adjacency relation captures all records of a single person across all silos simultaneously, and verify that the standard basic composition bound carries over to this adjacency model. Within this framework we prove that de-anonymization undergoes a phase transition at k* = Theta(log n / epsilon^2) (population size n, per-silo RR parameter epsilon): a Fano lower bound shows any estimator fails for k > k*. An explicit XOR + randomized-response construction demonstrates information synergy: each silo's output is individually uninformative about the target, yet the joint mutual information is strictly positive. For non-coordinated binary randomized-response mechanisms, we prove that de-anonymization is inevitable once k exceeds the threshold, establishing that cross-silo coordination is necessary. These results provide a baseline threat model and Theta-level threshold for cross-silo inference attacks under local DP.

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

Generation of Maximal Snake Polyominoes Using a Deep Neural Network

Maximal snake polyominoes are difficult to study numerically in large rectangles, as computing them requires the complete enumeration of all snakes for a specific rectangle size, which corresponds to a brute force algorithm. This hinders the study of maximal snakes in larger rectangles. Moreover, most enumerable snakes lie in small rectangles, obscuring large-scale patterns. In this paper, we investigate the contribution of a deep neural network to the generation of maximal snake polyominoes from a data-driven training, where the maximality and adjacency constraints are not encoded explicitly, but learned. To this extent, we experiment with a denoising diffusion model, which we referred as Structured Pixel Space Diffusion (SPS Diffusion). We find that SPS Diffusion generalizes from small rectangles to larger ones, generating valid snakes up to 28x28 squares and producing maximal snake candidates on squares close to the current computational limit. The model is, however, prone to errors such as branching, cycles, or multiple snake components. Overall, the diffusion model is promising and suggests that complex combinatorial objects can be understood by deep neural networks, which is useful in their investigation.

16.
arXiv (CS.CV) 2026-06-15

HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities

Reward models guide text-to-image (T2I) systems toward outputs aligned with human preferences. However, typical reward models such as HPSv3 are trained on pre-annotated data from earlier T2I models, without accounting for quality discriminative shifts arising from evolving model capabilities and reinforcement learning (RL) iterations, limiting their broader applicability. In this work, we propose HPSv3++, a reward model framework that elevates the HPSv3 model for varying T2I model capabilities and their RL iteration changes across the full capability-iteration spectrum. Specifically, we first introduce HPDv3++, a 212K dual-dimension preference dataset annotated for text fidelity and aesthetic quality using a recent high-capability (Qwen-Image) model with human supervision. We then propose a two-stage training framework. Stage 1 employs data-aware orthogonal gradient projection to incorporate diverse aesthetic perception from HPDv3++ while preserving the original effective human preference knowledge in HPSv3. Stage 2 further leverages unlabeled data from T2I models spanning different capability levels and RL iterations, and introduces a joint capability-iterations conditioned signal for the reward model together with a standard deviation-driven unsupervised guidance mechanism, strengthening reward model across the capability-iteration spectrum. HPSv3++ achieves state-of-the-art preference prediction, outperforming HPSv3 9.8% on HPDv3, 5.5% on GenAI-Bench, while achieving 79.1%/88.1% on our proposed HPDv3++. When used for T2I RL training, it consistently improves GenEval scores across diverse T2I models, demonstrating its wide-range capabilities. The code is available at https://github.com/PlantPotatoOnMoon/HPSv3-PlusPlus.

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

Chronological Thinking in Full-Duplex Spoken Dialogue Language Models

Recent advances in spoken dialogue language models (SDLMs) reflect growing interest in shifting from turn-based to full-duplex systems, where the models continuously perceive user speech streams while generating responses. This simultaneous listening and speaking design enables real-time interaction and the agent can handle dynamic conversational behaviors like user barge-in. However, during the listening phase, existing systems keep the agent idle by repeatedly predicting the silence token, which departs from human behavior: we usually engage in lightweight thinking during conversation rather than remaining absent-minded. Inspired by this, we propose Chronological Thinking, an on-the-fly conversational thinking mechanism that aims to improve response quality in full-duplex SDLMs. Specifically, chronological thinking presents a paradigm shift from conventional LLM thinking approaches, such as Chain-of-Thought, purpose-built for streaming acoustic input. (1) Strictly causal: the agent reasons incrementally while listening, updating internal hypotheses only from past audio with no lookahead. (2) No additional latency: reasoning is amortized during the listening window; once the user stops speaking, the agent halts thinking and begins speaking without further delay. Experiments demonstrate the effectiveness of chronological thinking through both objective metrics and human evaluations show consistent improvements in response quality. Furthermore, chronological thinking robustly handles conversational dynamics and attains competitive performance on full-duplex interaction metrics.

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

From Static Inference to Dynamic Interaction: A Survey of Streaming Large Language Models

Standard Large Language Models (LLMs) are predominantly designed for static inference with pre-defined inputs, which limits their applicability in dynamic, real-time scenarios. To address this gap, the streaming LLM paradigm has emerged. However, existing definitions of streaming LLMs remain fragmented, conflating streaming generation, streaming inputs, and interactive streaming architectures, while a systematic taxonomy is still lacking. This paper provides a comprehensive overview and analysis of streaming LLMs. First, we establish a unified definition of streaming LLMs based on data flow and dynamic interaction to clarify existing ambiguities. Building on this definition, we propose a systematic taxonomy of current streaming LLMs and conduct an in-depth discussion on their underlying methodologies. Furthermore, we explore the applications of streaming LLMs in real-world scenarios and outline promising research directions to support ongoing advances in streaming intelligence. We maintain a continuously updated repository of relevant papers at https://github.com/EIT-NLP/Awesome-Streaming-LLMs.

19.
arXiv (CS.CV) 2026-06-16

ResEdit: Residual embeddings for precise generative image editing

Conditional diffusion image generators can be repurposed for editing through inversion, without the need for large-scale paired fine-tuning data. However, producing high-quality, targeted edits while maintaining image identity and global consistency remains challenging, as weakly conditioned inversion often embeds conflicting image features into the noise. We demonstrate that incorporating a residual image encoding as additional conditioning enables both improved identity preservation and better editability. We optimize this residual encoding to provide a strong conditioning signal for reconstruction, thereby reducing the reliance on inversion and susceptibility to its aforementioned pitfalls. To ensure this residual does not interfere with desired edits, we incorporate a gradient reversal-based optimization strategy that disentangles the residual from the edited condition. We illustrate our method's ability to produce high-fidelity results across precise intrinsic-based editing and relighting, and show proof-of-concept text-guided manipulation.

20.
arXiv (CS.AI) 2026-06-12

A Mathematical Theory of Value: a synthesis on goal-directed agency under resource constraints

作者:

arXiv:2606.12502v1 Announce Type: cross Abstract: We propose that value – the quantity goal-directed agents create, destroy, and exchange – is a lawful structural quantity in the same category as information. Following Shannon's method, we make one ruthless abstraction: value is the rate at which an agent converts a resource into goal-progress, relative to a frame fixed by its goal. A scale-invariance axiom forces a logarithmic measure, $V=\sum_i k_i \ln e_i$; compounding of a reinvested resource forces the same form via the ergodicity argument of Peters (2019). The two routes are kin rather than independent; their agreement is a consistency check, not an over-determination. We derive a coding theorem of value: $\Delta G \le I(X;Y)$, achieved by Bayes-proportional allocation; realized value decomposes as $G=D(q\|r)-D(q\|p)$, identifying misalignment with measurable waste. For populations, value is frame-relative while price is frame-independent; a fleet that pools its resource and fuses its perception inherits the ceiling $G_{\mathrm{fleet}} \le I(X;Y_{1:m}) \le H(X)$ (a corollary; an earlier sum-form claim was wrong and is corrected in v5). A dynamical layer yields an is/ought asymmetry from which alignment emerges as a control-stability condition with a closed-form residual. We test the single-frame laws on live language models in a pre-registered scale-up: perception mutual information tracks realized capability rather than parameter count (Spearman $\rho = 0.977$ pooled over 30 model$\times$domain points), out-of-sample $\Delta G$ tracks $I(X;Y)$, and over-confidence is measurable dissipation; a further pre-registered test shows the bridge is shape-invariant across four task shapes ($n=42$, slope 0.953). None of the mechanisms is individually new – generalized Kelly, Armstrong & Mindermann (2018), classical control; the contribution is their unification and the governance mapping (incentive design over oversight) that follows.

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

Neural FOXP2 – Language Specific Neuron Steering for Targeted Language Improvement in LLMs

LLMs are multilingual by training, yet their lingua franca is often English, reflecting English language dominance in pretraining. Other languages remain in parametric memory but are systematically suppressed. We argue that language defaultness is governed by a sparse, low-rank control circuit, language neurons, that can be mechanistically isolated and safely steered. We introduce Neural FOXP2, that makes a chosen language (Hindi or Spanish) primary in a model by steering language-specific neurons. Neural FOXP2 proceeds in three stages: (i) Localize: We train per-layer SAEs so each activation decomposes into a small set of active feature components. For every feature, we quantify English vs. Hindi/Spanish selectivity overall logit-mass lift toward the target-language token set. Tracing the top-ranked features back to their strongest contributing units yields a compact language-neuron set. (ii) Steering directions: We localize controllable language-shift geometry via a spectral low-rank analysis. For each layer, we build English to target activation-difference matrices and perform layerwise SVD to extract the dominant singular directions governing language change. The eigengap and effective-rank spectra identify a compact steering subspace and an empirically chosen intervention window (where these directions are strongest and most stable). (iii) Steer: We apply a signed, sparse activation shift targeted to the language neurons. Concretely, within low to mid layers we add a positive steering along the target-language dominant directions and a compensating negative shift toward the null space for the English neurons, yielding controllable target-language defaultness.

22.
arXiv (CS.CL) 2026-06-16

VeriGraph: Towards Verifiable Data-Analytic Agents

LLM-based agents have demonstrated strong capabilities in data-intensive analytical tasks, yet their outputs are rarely verifiable: a reliance on linear text trajectories makes their reasoning difficult to audit. In particular, deterministic computations over raw data and semantic deductions over natural-language claims are often entangled in an unstructured stream, leaving numerical conclusions hard to reproduce and qualitative judgments hard to inspect. To address this, we propose VeriGraph, a traceable neuro-symbolic reasoning framework that enables agents to construct an explicit heterogeneous evidence directed acyclic graph (DAG) during execution. VeriGraph introduces three evidence-expansion primitives, namely computational, grounding, and derivational expansion, to connect raw data, interpreter variables, computed results, and natural-language claims in a unified graph. Under this formulation, structural traceability is reduced to graph reachability from raw data sources to terminal claims, while semantic support is measured by claim-level evidence evaluation. To improve graph construction, we further design a graph-based policy optimization strategy with a composite reward that jointly supervises answer correctness, computational integrity, and derivational coherence. Experiments on four benchmarks show that VeriGraph-8B achieves the highest overall score among all baselines. More importantly, VeriGraph produces auditable evidence graphs with substantially stronger claim grounding, achieving a 87.61\% Grounding Rate under our claim-level evidence support evaluation. These results suggest that explicit evidence-graph construction is a promising path toward verifiable data-analytic agents. Our code is available at https://github.com/ignorejjj/VeriGraph.

23.
arXiv (CS.CL) 2026-06-15

X-OPD: Cross-Modal On-Policy Distillation for Capability Alignment in Speech LLMs

While the shift from cascaded dialogue systems to end-to-end (E2E) speech Large Language Models (LLMs) improves latency and paralinguistic modeling, E2E models often exhibit a significant performance degradation compared to their text-based counterparts. The standard Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) training methods fail to close this gap. To address this, we propose X-OPD, a novel Cross-Modal On-Policy Distillation framework designed to systematically align the capabilities of Speech LLMs to their text-based counterparts. X-OPD enables the Speech LLM to explore its own distribution via on-policy rollouts, where a text-based teacher model evaluates these trajectories and provides token-level feedback, effectively distilling teacher's capabilities into student's multi-modal representations. Extensive experiments across multiple benchmarks demonstrate that X-OPD significantly narrows the gap in complex tasks while preserving the model's inherent capabilities.

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

Task-Aware Structured Memory for Dynamic Multi-modal In-Context Learning

Multi-modal large language models (MLLMs) depend on in-context learning (ICL) for rapid task adaptation, but their scalability is severely limited by finite context windows and the growing cost of key-value (KV) caches in long multi-modal sequences. Existing memory compression approaches typically rely on rigid token removal or sample-dependent importance estimation, which introduces bias, disrupts semantic structure, particularly for visual representations, and yields static memories that cannot adapt to new queries. We introduce TASM (Task-Aware Structured Memory), a training-free framework that addresses these limitations through task-aware, structure-preserving, and dynamically accessible memory construction. TASM employs task-vector guided compression to replace sample-specific signals with a task-level direction that captures shared relevance across demonstrations. To preserve the underlying manifold, it applies semantics-aware token merging via bipartite graph matching, aggregating tokens without destructive pruning. Finally, TASM structures memory into a hierarchy comprising a compact Core Memory and a Latent Bank, facilitating query-adaptive dynamic retrieval. Evaluations confirm TASM maintains high performance under heavy compression, effectively balancing efficiency with adaptability.

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

Muon$^p$: Muon with Fractional Spectral Powers

arXiv:2606.13867v1 Announce Type: new Abstract: Muon is an increasingly widely used optimizer that replaces a gradient $G=USV^\top$ with its polar factor $UV^\top$, thereby flattening the singular spectrum. However, full flattening discards singular-value information that may matter for adaptation. We introduce Muon$^p$, a Muon-style optimizer that instead uses fractional spectral-power updates $US^pV^\top$ for rational $p\in(0,1)$, interpolating between Muon and gradient descent. To make it practical, we prove that fractional spectral powers cannot be computed by any fixed univariate polynomial iteration, and furthermore derive low-degree odd bivariate recurrences that approximate $US^pV^\top$ using only matrix multiplications, preserving Muon's matrix-multiplication-only structure and compute complexity. We show that Muon$^p$ maximizes the linear improvement in loss under the Schatten $q$-norm for $q=1+\frac{1}{p}$. Empirically, Muon$^p$ is especially effective for finetuning: on billion-scale models, Muon$^p$ improves validation perplexity and downstream task performance. We further analyze when Muon$^p$ is less suitable, through the lens of spectral geometry. Our results reveal important insights on when preserving the singular spectrum can bring significant gains, and introduce a principled way to achieve them.