arXiv:2606.02392v1 Announce Type: new
Abstract: We ask whether dependency topology correlates with functional load-bearing organization as recoverable geometry -- not as a metaphor, but as a measurable structural property detectable by multilayer network analysis. Across seven independent substrates, we show that hub persistence and rank divergence under the Functional Proximity Law recover operational organization that domain experts describe as logic: axiomatic load-bearing structure in formal mathematics, control and contract structure in legacy software, conserved hub grammar across approx. 600 million years of neural evolution, catalytic role organization in a published prebiotic autocatalytic network, carry-path dominance in a 4-bit digital circuit, betweenness persistence in the ISCAS85 c432 standard benchmark (n=196), and a directional formal-systems replication in the Coq Corelib (n=17). A key methodological finding: degree-based hub persistence is weak between physical wiring and simulation state-correlation layers (r=0.21 in c432), while betweenness-based persistence is stronger (r=0.77 in the 4-bit ALU post-hoc; r=0.34 in c432). The ISCAS85 pre-registered primary hypothesis was CONFIRMED (degree r=0.426, p=0.002, Spearman r=0.551). The formal-systems claim is supported by two proof-assistant corpora: Lean 4 mathlib4 (CONFIRMED, r=0.777, p=0.004) and Coq Corelib (PARTIAL, direction confirmed, r=0.288, p=0.287, n=17, underpowered). All seven experiments were pre-registered before analysis.
Science Journals
arXiv:2606.01165v1 Announce Type: cross
Abstract: Ramesh's 2023 dissertation introduces the categorical notions of introspective theories and geminal categories, which formalize "self-internalizing" structures sharing the form of L\"ob's theorem ($\Box A \vdash A$ implies $\vdash A$). We reorganize the theory of geminal categories in a self-contained manner by introducing "code structures on fibrations," which serve as a categorical abstraction of G\"odel coding. This framework leads to a significant simplification of the proof of L\"ob's theorem for geminal categories, as well as to a new categorical counterpart of the G\"odel-L\"ob axiom ($\Box(\Box A \to A) \to \Box A$). This formulation offers an accessible framework for Ramesh's approach and suggests connections to modal type theories, where similar meta- and object-level interactions arise.
arXiv:2606.00682v1 Announce Type: cross
Abstract: Layered halide perovskites host exceptionally strong excitons, whose optical signatures are usually interpreted as absorptive resonances on a smooth dielectric background. Strong excitons, however, can also reshape the dielectric response itself and drive the real permittivity negative, opening a reflective band: the visible, excitonic analogue of an infrared Reststrahlen band. Whether bare (PEA)$_2$PbI$_4$ slabs reach this regime has remained unclear. Here we show that low-temperature transmission of (PEA)$_2$PbI$_4$ slabs, driven by the intralayer-exciton manifold, evolves with increasing thickness from an excitonic dip into a broad near-zero-transmission interval with compressed Fabry-P\'erot-like fringes. Transfer-matrix analysis with an effective Lorentz-oscillator dielectric response reproduces this crossover, reconstructs a finite negative-Re($\varepsilon$) window, and implies near-ultrastrong exciton-photon coupling. Calculated field maps show suppressed in-plane field penetration within this interval and a driven longitudinal response near the high-energy ($\varepsilon=0$) edge. These results identify (PEA)$_2$PbI$_4$ slabs as a cavity-free visible-frequency excitonic Reststrahlen material.
arXiv:2606.00799v1 Announce Type: cross
Abstract: A classic theorem of Weyl (1921) states that a Weyl metric -- a natural generalisation of a pseudo-Riemannian metric -- is uniquely determined by its conformal and projective structures (i.e. by its conformal structure and its set of unparametrised geodesics). An equivalent formulation of Weyl's result is that a torsion-free linear connection compatible with a pseudo-Riemannian conformal structure is uniquely determined by its projective structure. We discuss analogous results for suitably defined notions of conformal structure for Galilei and Carroll geometry, i.e. for spacetime geometries arising as the `non-relativistic' and `ultra-relativistic' limits of Lorentzian geometry.
arXiv:2606.01479v1 Announce Type: new
Abstract: Integrating large language models (LLMs) into text-to-speech (TTS) systems has improved speech expressiveness, yet interpretable emotional control remains challenging. Existing approaches primarily rely on external conditioning or global activation steering, offering limited insight into the internal representations underlying emotional control. In this work, we analyze emotion-related variation in the semantic hidden states of LLM-based TTS models using sparse autoencoders (SAEs) to identify sparse latent features. Our analysis shows that emotional variation is distributed across multiple sparse latent features, while intervening on a small subset enables interpretable emotion control. Building on this observation, we introduce a feature-level intervention framework for bidirectional emotion induction and suppression without modifying backbone parameters. We further show that distinct latent features are associated with specific acoustic attributes (e.g., pitch), suggesting that emotional expression arises from coordinated latent contributions rather than a single global shift. Empirically, steering these sparse latent features achieves comparable or superior emotion induction and suppression performance relative to global steering and existing TTS baselines.
arXiv:2606.02167v1 Announce Type: new
Abstract: Engineers designing production systems need to verify that a given layout supports all required production sequences. Automated planning techniques can answer such questions, but formulating the required planning problems in the Planning Domain Definition Language (PDDL) demands specialized expertise that production engineers typically lack. Asset Administration Shells (AAS) have emerged as the standardized Digital Twin for industrial assets in Industry 4.0. We show that AAS capability models, structured using four established Industry 4.0 standards (VDI 3682 for process descriptions, IEC 61360-1 for semantic property qualification, IDTA 02011 for type hierarchies, and IDTA 02016 for instance descriptions), contain sufficient information to generate complete PDDL problems automatically. Unlike prior work that introduced PDDL-specific submodels, our approach derives all planning elements from domain-level descriptions of resource functions, so-called capabilities, allowing engineers to model capabilities without any exposure to PDDL syntax or planning concepts. Our extraction algorithm transforms distributed Multi-AAS architectures into complete PDDL planning problems. We validate the approach on AAS models of a laboratory production system, comparing four layout variants using optimal planning to demonstrate how engineers can systematically explore design trade-offs by modifying the AAS model and regenerating the planning domain
arXiv:2606.02172v1 Announce Type: new
Abstract: Learning discriminative visual representations from distributed, heterogeneous data is a fundamental challenge in Federated Learning (FL). Prototype-based methods address statistical heterogeneity by sharing class-level representations across clients but create a distance-dependent gradient pressure that is particularly severe during early training rounds: alignment pressure applied to immature global prototypes, aggregated from noisy local representations, generates large gradients that suppress the emergence of local discriminative structure. The result is a poorly organized embedding space and degraded recognition performance, particularly under severe non-IID conditions. We propose FedSAP, a framework that stabilises federated representation learning through two complementary mechanisms: a deterministic alignment curriculum that delays global alignment until local representations become stable and a geometry-driven proxy separation loss that enforces inter-class structure on the unit hypersphere using the existing prototype bank without introducing additional parameters or communication overhead. Together, these mechanisms produce compact, well-separated class clusters without altering the underlying communication protocol between federation's participants. Experiments across three benchmarks and varying degrees of heterogeneity show gains of up to 4 percentage points over the prototype-based baselines evaluated, with improvements most pronounced under high heterogeneity. The representational nature of our framework further enables a straightforward extension to semi-supervised settings, where unlabelled data is incorporated with minimal modification, underscoring the generality of scheduled alignment as a design principle.
arXiv:2606.01840v1 Announce Type: new
Abstract: Many difficult search problems cannot be solved by algorithms such as A* using only RAM. Search algorithms which use external memory such as SSDs and HDDs with much higher capacity than RAM have been proposed in previous work, but previous work has focused on delayed duplicate detection approaches, as well as complex immediate duplicate detection (IDD) methods, and relatively simple methods for IDD have not been systematically studied. In addition, the effect of OS-level mechanisms for managing and speeding up accesses to external memory, such as page caches, has not been studied. This paper addresses these gaps in the literature by evaluating and analyzing the performance of simple baseline approaches for IDD-based A*.
arXiv:2606.01846v1 Announce Type: new
Abstract: Mosquito-borne infectious diseases cause more than 700000 deaths worldwide each year. The long-term use of conventional chemical insecticides has induced serious resistance problems, creating an urgent need to develop novel, highly effective, and ecologically sustainable alternatives. While existing artificial intelligence approaches in this domain have focused primarily on activity prediction and classification, they leave a critical gap in the de~novo generation of novel molecular scaffolds. In this study, we propose Mos-Gen, a motif-aware generative collaborative framework that couples the pretrained molecular representation model Uni-Mol with a variational autoencoder (VAE), specifically tailored for the design of disulfide-containing allicin derivatives as mosquito insecticides. Among the generated candidates, fourteen compounds -- comprising nine predicted positives and five predicted negatives -- were selected for chemical synthesis and experimental validation. The hit rate among the predicted positives reached 78%, whereas none of the predicted negatives exhibited mosquitocidal activity. These experimental results fully validated the high-precision screening capability of the Mos-Gen framework.
arXiv:2606.01847v1 Announce Type: new
Abstract: Diffusion-based Vision-Language-Action policies achieve remarkable success in robotic manipulation, yet commit a fundamental geometric error we term the $\textbf{Euclidean Fallacy}$: representing SE(3) poses as flat $\mathbb{R}^{12}$ vectors. This approximation induces (1) manifold drift violating SO(3) constraints, (2) broken equivariance under coordinate transformations, and (3) non-geodesic trajectories with excessive kinematic cost. We introduce $\textbf{Lie Diffuser Actor (LDA)}$, a diffusion framework operating intrinsically on SE(3). Our method injects noise through left-invariant SDEs, predicts scores in the tangent space, and retracts samples via the exponential map. This formulation eliminates manifold drift by construction while guaranteeing coordinate-frame equivariance and geodesic optimality. On CALVIN ABC$\rightarrow$D, LDA improves average task length from $3.27$ to $3.51$ ($+7.3\%$). We further validate our method on real robot and the results show that our methodology outperforms the baseline on majority tasks.
arXiv:2606.02365v1 Announce Type: new
Abstract: Shampoo is attracting considerable attention for its superior performance on large-scale optimization benchmarks; yet it faces a significant practical bottleneck: the prohibitive computational overhead of matrix inversion. To mitigate this, practitioners typically rely on stale preconditioner updates, creating a fundamental trade-off between computational efficiency and optimization fidelity. In this work, we provide a theoretical study of staleness through the complementary lenses of convergence and stability. While staleness improves computational efficiency, it inherently degrades performance and introduces numerical instability. Crucially, we identify that damping, acting as a numerical stabilizer, can effectively suppress these negative effects. Guided by this analysis, we propose FOAM, an adaptive algorithm that stabilizes training by dynamically controlling both the damping factor and the eigendecomposition frequency based on an approximation of the staleness-oriented error. Experimental results demonstrate that FOAM reduces wall-clock time compared to standard Shampoo while maintaining robust convergence.
arXiv:2606.02487v1 Announce Type: new
Abstract: Effective "all-team" summarization in high-complexity settings like the Neonatal Intensive Care Unit (NICU) requires aggregating insights from diverse disciplines (physicians, nurses, therapists) spread across hundreds of clinical free-text notes. Simply pooling heterogeneous text often leads to incoherent outputs. Structured summarization therefore first requires accurate categorization of sentence-level provenance across multi-source notes. This pilot study introduces a clinical provenance categorization pipeline using supervised fine-tuning (SFT) of large language models (LLMs). We adapted two Llama-3 models (8B and 70B) to MedSecId, a corpus of 2,002 MIMIC-III (Adult ICU) notes annotated with clinical provenance headers, achieving in-domain Macro F1 scores above 92% for both models. To evaluate cross-domain generalization, we assessed model capacity (8B vs. 70B) and quantization on a gold-standard dataset of 227 sentence-level spans derived from three multi-disciplinary NICU summaries. Experimental results demonstrate a scale-dependent transfer effect: while SFT produced only marginal changes for the 8B model, it substantially improved the 70B model, increasing Macro F1 by 7%. Notably, the quantized fine-tuned 70B model outperformed its full-precision baseline while substantially reducing computational requirements. These findings suggest that sufficient model capacity is critical for preserving semantic flexibility during cross-domain clinical transfer and that efficient quantized adaptation can enable structured provenance modeling for downstream summarization.
arXiv:2606.01848v1 Announce Type: new
Abstract: Search-and-rescue (SAR) requires embodied agents to explore unfamiliar environments under multimodal uncertainty, perform multi-stage interactions, and retrieve spatial memory over long horizons. Existing benchmarks typically evaluate these capabilities in isolation, leaving unclear how failures compound when they must be composed in realistic workflows. We introduce RescueBench, a photo-realistic diagnostic benchmark that instantiates SAR as a four-stage pipeline: multimodal exploration, target rescue, memory-guided return, and final handoff. By combining sequential task composition with stage-level evaluation, RescueBench enables analysis of how exploration and memory failures propagate through embodied rescue workflows. It contains five progressive difficulty levels that vary in environmental complexity, clue ambiguity, and spatial hierarchy, along with an automatic episode generation and annotation pipeline for scalable evaluation and training. We evaluate seven baselines, an oracle reference, and human players, showing that no baselines complete the full task at the greatest difficulty. Stage-level diagnosis identifies autonomous exploration as the dominant failure mode and spatial memory as a second, independent bottleneck, suggesting that these limitations are not resolved by current topological visual-language navigation or map-based methods. Code is available in https://github.com/wukui-muc/RescueBench
arXiv:2606.01873v1 Announce Type: new
Abstract: LLM-as-Aligner has emerged as a prevalent pre-training paradigm for Text-Attributed Graphs(TAGS), aligning graph and text modalities into a shared embedding space via CLIP-style contrastive learning. While effective on individual downstream tasks, we observe severe catastrophic forgetting when such models are sequentially fine-tuned on streaming tasks. Although parameter-efficient fine-tuning alleviates forgetting to some extent, it remains insufficient to resolve task interference and ineffective knowledge transfer. In this work, we study graph continual learning for LLM-as-Aligner models on TAGs, with the goal of mitigating interference while promoting positive transfer across tasks. This setting introduces two fundamental challenges: (1) heterogeneous downstream tasks induce shifting optimization objectives, hindering unified fine-tuning; and (2) graph and text encoders exhibit different sensitivities to adaptation, making uncoordinated updates prone to misalignment. To address these challenges, we propose G2LoRA, a continual learning framework for TAGs. G2LoRA unifies node-, link-, and graph-level tasks under a single graph--text alignment objective, and enables consistent optimization across domain/class/task incremental modes. To reduce task interference while encouraging positive transfer, G2LoRA performs category-aware gradient projection in structured subspaces, resolving conflicting updates and enabling conditional backward transfer to balance forward and backward knowledge flow. To further prevent cross-modal drift, G2LoRA introduces gradient magnitude modulation to coordinate update rates between graph and text encoders. Extensive experiments on benchmark datasets demonstrate that G2LoRA consistently outperforms strong baselines across different backbone architectures, achieving superior continual performance and transferability.
arXiv:2606.01577v1 Announce Type: new
Abstract: Methane is a major driver of near-term climate change, and rapidly identifying its emission sources is a critical climate intervention. Spaceborne hyperspectral imagery is the primary tool for this task, but the volume of data produced by each sensor makes ground-based detection impractical and necessitates onboard detection. Classical methods incur prohibitive computational cost on onboard hardware, while deep learning models are fast but fall short on detection quality. We propose FLAME, a physics-guided neural operator that builds the physics of methane absorption directly into its architecture. On the methane detection benchmark, FLAME achieves the highest detection accuracy among all evaluated methods, reduces the pixel-level false positive rate by nearly $3\times$ over the strongest neural baseline, uses the fewest parameters among learned baselines, and runs within the latency budget of onboard satellite hardware.
arXiv:2606.01328v1 Announce Type: cross
Abstract: We have developed a fully Bayesian survival-analysis framework that reformulates inference about system lifetimes in terms of hazard and survival functions, and extends this representation to interacting actors. Starting from J.~Richard Gott's Copernican principle, we express the scale-free prior as a baseline hazard $\lambda(t)=1/t$, thereby linking a static prior over lifetimes to the dynamic language of survival analysis. In this formulation, Bayesian updating corresponds to conditioning on survival, while the resulting posterior distribution admits a natural representation in terms of hazard and survival functions. The approach is intended for settings where data are sparse or unreliable, and where a scale-free, assumption-light baseline is preferable to heavily parameterized models.
Building on this foundation, we derive general expressions for two-actor systems that characterize joint survival, conditional lifetimes, and comparative outcomes without requiring a specific parametric form of interaction. This yields a flexible and modular framework in which baseline dynamics are separated from interaction effects, allowing different mechanisms to be incorporated transparently. Thus, the primary contribution is a general hazard-based formulation of Bayesian updating and its extension to interacting systems
To illustrate the framework, we consider a multiplicative resource-depletion specification in which interaction modifies the baseline hazard through cumulative engagement intensity. This example demonstrates how interaction terms can be embedded while preserving analytical tractability, including closed-form expressions under simplifying assumptions. We further provide a stylized application to an asymmetric two-actor conflict, the 2026 US/Israel--Iran hostilities, to highlight the qualitative implications of the approach.
arXiv:2606.02286v1 Announce Type: cross
Abstract: The deterministic incompressible Navier-Stokes equations are physically incomplete: any viscous fluid at finite temperature must exhibit thermal fluctuations whose form is dictated by the fluctuation-dissipation relation. We formulate the stochastic Navier-Stokes equations with the kinematically selected deformation Laplacian on compact Riemannian manifolds with strictly negative Ricci curvature. The fluctuation-dissipation relation, derived from a topological (Poincar\'e lemma) argument, uniquely determines the noise from the viscous operator. For the spectrally truncated system, we prove that the unique stationary distribution is the Gibbs measure (Gaussian in the mode amplitudes, because the nonlinear convective terms preserve energy), and that convergence to equilibrium is exponentially fast with rate at least $2\nu\lambda_\Def$, where $\nu$ is the kinematic viscosity and $\lambda_\Def$ is the spectral gap of the deformation Laplacian. The spectral gap satisfies $\lambda_\Def \geq \kappa^2$ when $\Ric \leq -\kappa^2 g$, and is independent of the volume of the domain. On flat space, the analogous thermalisation rate vanishes in the infinite-volume limit. The equilibrium velocity-velocity correlation function decays exponentially in geodesic distance, in contrast to the algebraic decay on flat space. These results provide a rigorous statistical-mechanical foundation for viscous fluids on negatively curved manifolds and illustrate how the geometry of the domain controls not only the deterministic dynamics but also the approach to thermal equilibrium.
arXiv:2606.02305v1 Announce Type: cross
Abstract: Understanding how speech foundation models relate to human cortical activity is a key challenge for computational neuroscience. Here, we investigate how internal representations from Whisper predict intracranial ECoG responses during naturalistic speech perception. We introduce a time-resolved neural encoder that combines speech embeddings with a recurrent temporal model and soft attention, allowing us to examine layer-wise brain alignment. Intermediate Whisper layers provide the strongest correspondence with neural activity, supporting a hierarchical match between model representations and cortical speech processing. Comparisons with baselines show that high-resolution ECoG responses benefit from temporally structured modelling beyond linear mappings from the same speech representations. In addition, attention maps reveal temporally local alignment between speech embeddings and neural responses, while a phonemic interpretability analysis identifies anatomically coherent phoneme-category organization among encoding-informative electrodes. Together, these results suggest that speech foundation models offer a useful framework for studying time-resolved cortical speech representations.
arXiv:2606.01865v1 Announce Type: new
Abstract: Diffusion policies have recently emerged as a powerful framework for robotic manipulation. However, like other behavior cloning methods, they remain vulnerable to distributional shift, often requiring human-in-the-loop interventions to correct failures during deployment. These interactions naturally provide paired supervision in the form of the robot's undesired actions and the human teacher's corrective actions. Yet existing data aggregation pipelines and standard behavior cloning losses largely ignore this negative signal from undesired actions, leading to overfitting to teacher's actions and an increasing reliance on costly expert data. To address this limitation, we propose Set-Supervised Diffusion Policy (SDP), a novel learning framework that utilizes contrastive action-chunk data to train diffusion policies from human corrections. From paired positive and negative action-chunks, SDP constructs a set of desired action-chunks and designs a training pipeline that encourages the diffusion policy to align with the set. Through extensive experiments across multiple robotic manipulation tasks, we demonstrate that SDP consistently improves policy performance, with particularly strong gains in robustness to noisy data. Moreover, SDP induces high-quality aggregated datasets, enabling more efficient and reliable policy learning from human-in-the-loop corrections. Our code is available at https://set-supervised-diffusion-policy.github.io/.
arXiv:2606.01885v1 Announce Type: new
Abstract: With the evolution of generative models, deepfakes have achieved near-perfect semantic realism, leaving forensic traces only in subtle structural anomalies. However, existing single-view paradigms often fail to generalize, as dominant semantic features overwhelm subtle artifact cues within entangled representations. This imbalance leads to overconfident yet brittle predictions -- a phenomenon we term the Semantic Masking Effect. To address this challenge, we propose a reliable framework called Divide-and-Conquer Multi-View Evidential Learning (DiCoME) for Deepfake Detection. In the "Divide" phase, we employ Geometric View Purification to decompose the entangled representation space through principled geometric projection. This process suppresses semantic interference within artifact-sensitive representations, forming the foundation for decorrelated yet complementary semantic and artifact views. In the "Conquer" phase, we leverage Uncertainty-Aware Evidential Learning to synthesize these distinct views. By explicitly modeling the "epistemic conflict" between semantic and artifact cues, this mechanism provides calibrated uncertainty estimates instead of forcing rigid deterministic decisions. Extensive experiments across multiple benchmarks demonstrate that our method consistently outperforms existing approaches in generalization performance, while providing reliable uncertainty estimation for trustworthy deepfake detection. Code is available at https://github.com/kxl0825/DiCoME.git.
arXiv:2606.01636v1 Announce Type: new
Abstract: Post-training via Group Relative Policy Optimization (GRPO) has emerged as a powerful paradigm for aligning flow-based generative models with human preferences. However, the iterative denoising nature of flow models incurs substantial costs when generating group rollouts for policy-gradient updates, compelling existing methods to train with extremely few denoising steps. This temporal sparsity severely restricts preference optimization: reward feedback can only reach a handful of stages per trajectory, leaving the vast majority of intermediate denoising steps without direct supervision and thus compromising alignment granularity. To address this, we propose Pave-GRPO, which reformulates the GRPO objective through Principled average velocity decomposition. Rather than generating expensive high-step rollouts, we maintain efficient few-step group sampling but decompose each coarse transition into an equivalent ensemble of finer sub-trajectories spanning multiple intermediate timesteps. This propagates reward feedback to a denser set of temporal stages for more comprehensive preference alignment without additional generation cost. This design offers two benefits: (i) zero-cost horizon expansion: through the direct reuse of piece-wise group samples and their associated rewards, Pave-GRPO significantly broadens the effective optimization scope under fixed sampling budgets; and (ii) comprehensive temporal supervision: by equivalently decomposing an instantaneous velocity target into a multi-timestep ensemble, it distributes reward signals across more intermediate stages of the denoising process, enabling finer-grained and more thorough preference optimization. Extensive experiments validate that Pave-GRPO effectively advances preference alignment across different reward settings, offering comprehensive performance enhancement.
arXiv:2606.01747v1 Announce Type: new
Abstract: Through digital humanities research and scale-up historical data analysis, a significant amount of traditional historical text is converted into structured knowledge graphs. This paper provides a high-level architecture that combines bidirectional encoder representations of transformers (BERT) and graph neural networks (GNN) to extract the entities and relationships from various types of historical texts. The texts of traditional history resolve linguistic ambiguities, references limited by context, and a lack of established grammatical norms in a systematic way. This study develops a new image retrieval system based on FastRQNet and pre-trained vision-language model Vilt-qaformer+RoBInet in accordance with the aforementioned recommendations. The experiments make full use of a comprehensive collection of municipal records, parliamentary documents, and historical correspondence. When compared to conventional rule-based techniques and other popular deep-learning baselines, the joint BERT-GNN system obtains greater Precision, Recall, and F1-score (Table 2). Complex nested structures and implicit reference issues can be handled by this structure with sufficient accuracy and thoroughness when creating knowledge graphs. The aforementioned experiments show that combining relational graph learning algorithms with context-sensitive semantic representation techniques can automatically extract historical data to add accumulated wisdom to the knowledge repository.
arXiv:2606.02375v1 Announce Type: new
Abstract: We evaluate whether compact domain-specialized ASR models can outperform massively multilingual foundation models for conversational African speech across 19 languages in the WAXAL corpus. Fine-tuned edge models achieve a macro-averaged WER of $38.0\%$ compared to $64.9\%$ for the best zero-shot baseline, a $26.9$ percentage-point reduction using models $3-40\times$ smaller. Results confirm that domain specialization dominates scale for spontaneous African speech. Cross-domain evaluation shows that fine-tuned models recover usable performance on out-of-distribution (OOD) speech, while zero-shot models regain an advantage when the test domain matches their pretraining distribution. A distributed native-speaker audit across all surveyed languages produces a linguistically-grounded error taxonomy, showing that CTC and autoregressive architectures behave differently across language families. We further show that WER alone misrepresents performance for syllabary-script languages where CER/WER ratios reveal substantially higher character-level accuracy than headline WER suggests. Finally, to contribute to future African ASR research, we release all model weights, fine-tuning and evaluation scripts, and a cleaned WAXAL subset covering all $19$ languages.
arXiv:2606.01934v1 Announce Type: new
Abstract: Large language models achieve remarkable performance via extended chain-of-thought (CoT) reasoning, yet this lengthy process incurs substantial inference overhead. Existing CoT compression methods struggle with inflexible manual length budgets, computationally expensive multi-stage training pipelines, and fragile scalability restricted to small models. We propose HMPO (Hybrid Median-length Policy Optimization), a cost-effective, single-stage reinforcement learning framework. HMPO efficiently compresses CoT via three synergistic components: an adaptive median-based budget derived from successful rollouts to eliminate manual tuning, a cosine-decay token reward for smooth length penalization, and a multiplicative reward formulation that substantially mitigates trivial reward hacking by strictly prioritizing answer correctness. Trained exclusively on mathematical data, HMPO generalizes seamlessly across math, code, science, and instruction-following tasks. Extensive experiments scaling from 9B to 122B parameters across dense and Mixture-of-Experts (MoE) architectures demonstrate that HMPO achieves 19%--46% token compression with negligible accuracy degradation, all while drastically reducing training costs compared to existing multi-stage baselines.
arXiv:2510.08948v4 Announce Type: replace
Abstract: Effective e-commerce risk management requires in-depth case investigations to identify emerging fraud patterns in highly adversarial environments. However, manual investigation typically requires analyzing the associations and couplings among multi-source heterogeneous data, a labor-intensive process that limits efficiency. While Large Language Models (LLMs) show promise in automating these analyses, their deployment is hindered by the complexity of risk scenarios and the sparsity of long-tail domain knowledge. To address these challenges, we propose Sherlock, a framework that integrates structured domain knowledge with LLM-based reasoning through three core modules. First, we construct a domain Knowledge Base (KB) by distilling structured expertise from heterogeneous knowledge sources. Second, we design a two-stage retrieval-augmented generation strategy tailored for case investigation, which combines input contextual augmentation with a Reflect & Refine module to fully leverage the KB for improved analysis quality. Finally, we develop an integrated platform for operations and annotation to drive a self-evolving data flywheel. By combining real-time hotfixes through KB updates with periodic logic alignment via post-training, we facilitate continuous system evolution to counteract adversarial drifts. Online A/B tests at JD dot com demonstrate that Sherlock achieves an 82% Expert Acceptance Rate (EAR) and a 386.7% increase in daily investigation throughput. An additional 90-day evaluation shows that the flywheel successfully recovers from performance decay caused by changing tactics twice, raising the EAR ceiling by around 3.5% through autonomous model updates.