arXiv:2605.17799v1 Announce Type: new
Abstract: Long-tailed out-of-distribution (LT-OOD) detection is often addressed with specialized training, including auxiliary out-of-distribution (OOD) data, abstention heads, contrastive objectives, energy losses, or gradient-conflict control. We show that these training mechanisms can obscure a simpler issue: frozen long-tailed representations may already contain useful OOD evidence, but raw Mahalanobis distance is distorted by frequency-coupled feature radius and poorly supported tail covariance. We propose Hyperspherical Pooled Mahalanobis (HPM), a post-hoc detector that normalizes features onto the unit sphere and replaces class-specific covariance with a pooled, ridge-regularized metric while keeping class means as semantic anchors. In CIFAR-LT experiments and an ImageNet-100-LT near-OOD boundary analysis, HPM improves raw Mahalanobis scoring; for Prior-Calibrated ERM (PC-ERM), it raises AUROC from 46.49 to 85.67 on CIFAR-10-LT and from 50.40 to 78.35 on CIFAR-100-LT. This simple PC-ERM+HPM pipeline also achieves the best Log Efficiency Score (LES; 3.08) on CIFAR-100-LT, retaining roughly 95% of the best CIFAR-100-LT AUROC observed among the compared post-hoc scores at substantially lower training-time cost. These results argue for evaluating representation quality, detector geometry, and training complexity as separate factors in LT-OOD detection.
Science Journals
arXiv:2605.17800v1 Announce Type: new
Abstract: Rearranging densely packed tabletop objects is challenging when parallel-gripper picks are infeasible without sufficient clearance around an object. This work studies the problem characteristics for practically motivated settings with uniformly sized blocks placed at planar tabletop grid locations. Since purely prehensile removal can become infeasible, a directional knock primitive is therefore introduced and the optimal knock-pick variant of the problem is formulated. The work proposes a series of abstractions wherein minimal constraining gadgets are covered to identify the necessary knocks. Utilizing a maximum-weight perfect matching on a graphical abstraction yields efficient polynomial-time computation of the optimal plan that minimizes the number of actions. Experiments are reported for increasing grid sizes in synthetic settings as well as in IsaacSim. The theoretical observations provide a promising stepping stone towards rigorously building efficient manipulation strategies that interleave prehensile and non-prehensile actions.
arXiv:2604.01658v2 Announce Type: replace
Abstract: Large language model (LLM)-based evolution is a promising approach for open-ended discovery, where progress requires sustained search and knowledge accumulation. Existing methods still rely heavily on fixed heuristics and hard-coded exploration rules, which limit the autonomy of LLM agents. We present CORAL, the first framework for autonomous multi-agent evolution on open-ended problems. CORAL replaces rigid control with long-running agents that explore, reflect, and collaborate through shared persistent memory, asynchronous multi-agent execution, and heartbeat-based interventions. It also provides practical safeguards, including isolated workspaces, evaluator separation, resource management, and agent session and health management. Evaluated on diverse mathematical, algorithmic, and systems optimization tasks, CORAL sets new state-of-the-art results on 10 tasks, achieving 3-10 times higher improvement rates with far fewer evaluations than fixed evolutionary search baselines across tasks. On Anthropic's kernel engineering task, four co-evolving agents improve the best known score from 1363 to 1103 cycles. Mechanistic analyses further show how these gains arise from knowledge reuse and multi-agent exploration and communication. Together, these results suggest that greater agent autonomy and multi-agent evolution can substantially improve open-ended discovery. Code is available at https://github.com/Human-Agent-Society/CORAL.
arXiv:2605.18690v1 Announce Type: new
Abstract: Integrated photonics has enabled a wide class of chip-scale light sources and quantum technologies. Within this field, microresonator-based degenerate optical parametric oscillators (DOPOs) have gained prominence. Above a critical power threshold, these systems undergo spontaneous symmetry breaking to settle into one of two stable, {\pi}-phase-shifted states -- a mechanism successfully used for quantum random number generation and photonic Ising machines. Here, we show that DOPOs based on the Kerr nonlinearity host a significantly broader range of nonlinear dynamics than previously explored. Using a silicon nitride microring resonator, we experimentally identify Hopf bifurcations that trigger a transition from stationary operation to self-sustained oscillations at MHz frequencies. By adjusting pump detunings and powers, we achieve turnkey control over these oscillatory regimes, navigating the system between stable binary states and periodic limit cycles. Furthermore, we report the experimental observation of period-doubling bifurcations, which numerical simulations reveal as the precursor to a cascading instability culminating in chaos at elevated pump powers. Our results establish a framework for controlling nonlinear instabilities in chip-scale parametric oscillators, with applications in programmable photonic hardware and dynamical optical computing.
arXiv:2411.15361v3 Announce Type: replace
Abstract: In the design of cellular manufacturing systems (CMS), numerous technological and managerial decisions must be made at both the design and operational stages. The first step in designing a CMS involves grouping parts and machines. In this paper, four integer programming formulations are presented for grouping parts and machines in a CMS at both the design and operational levels for a generalized grouping problem, where each part has more than one process plan, and each operation of a process plan can be performed on more than one machine. The minimization of inter-cell and intra-cell movements is achieved by assigning the maximum possible number of consecutive operations of a part type to the same cell and to the same machine, respectively. The suitability of minimizing inter-cell and intra-cell movements as an objective, compared to other objectives such as minimizing investment costs on machines, operating costs, etc., is discussed. Numerical examples are included to illustrate the workings of the formulations.
arXiv:2605.17345v1 Announce Type: new
Abstract: The release of public 3D medical image segmentation (MIS) datasets accelerates clinical research but simultaneously heightens risks of unauthorized AI model training. While Unlearnable Examples (UE) offer protection by injecting imperceptible perturbations to prevent effective model learning, existing methods primarily target 2D scenarios. They neglect the volumetric spatial correlations and inter-slice anatomical consistency inherent in 3D medical volumes, which serve as critical learning priors for 3D segmentation networks. To bridge this gap, we propose VoxShield, a UE framework that explicitly targets the volumetric inductive biases of 3D networks. Our core insight is that by systematically dismantling the cross-slice continuity that 3D architectures rely on, we can fundamentally impair their spatial aggregation process. Specifically, we introduce an Inter-Slice Frequency Consistency Disruption mechanism that maximizes the spectral divergence between adjacent slices, injecting structural incoherence along the $z$-axis. Complementing this structural attack, a Semantic Prediction Disruption module is incorporated. By maximizing the $\ell_1$ divergence between clean and perturbed logits, it forces the injected noise to penetrate the entire network and corrupt the final semantic mapping. Experiments on BraTS19 and FLARE21 demonstrate that VoxShield successfully degrades 3D segmentation performance, reducing the DSC from 80.0% to near 0.0% and from 88.6% to 6.8%, respectively. All protections are achieved with minimal perturbation ($\epsilon=4/255$) to preserve high visual fidelity. The code is available at https://github.com/KK266299/VoxShield.
arXiv:2510.04309v3 Announce Type: replace
Abstract: Controlling the behaviors of large language models (LLM) is fundamental to their safety alignment and reliable deployment. However, existing steering methods are primarily driven by empirical insights and lack theoretical performance guarantees. In this work, we develop a control-theoretic foundation for activation steering by showing that popular steering methods correspond to the proportional (P) controllers, with the steering vector serving as the feedback signal. Building on this finding, we propose Proportional-Integral-Derivative (PID) Steering, a principled framework that leverages the full PID controller for activation steering in LLMs. The proportional (P) term aligns activations with target semantic directions, the integral (I) term accumulates errors to enforce persistent corrections across layers, and the derivative (D) term mitigates overshoot by counteracting rapid activation changes. This closed-loop design yields interpretable error dynamics and connects activation steering to classical stability guarantees in control theory. Moreover, PID Steering is lightweight, modular, and readily integrates with state-of-the-art steering methods. Extensive experiments across multiple LLM families and benchmarks demonstrate that PID Steering consistently outperforms existing approaches, achieving more robust and reliable behavioral control. The code is publicly available at: https://github.com/dungnvnus/pid-steering
arXiv:2605.17801v1 Announce Type: new
Abstract: A measure for the complexity of a differentiable function f(x) on an interval is introduced. It is based on approximations of the function by piecewise constant functions. The measure takes into account the quality of the approximation and the number of intervals in the approximating function.
This measure, called the V-complexity of f(x), is shown to formalize some intuitions about the simplicity or complexity of f(x).
The V-complexity is then compared to another measure of complexity, namely how compressible an approximation of f(x) is. It is hypothesized that V-complexity is equivalent to the compression measure, in the case of the Run Length Encoding and the Lempel Ziv 77 algorithms.
V-complexity can be used as an ingredient in the definition of the Effective Complexity (EC) of a Complex System. When the perceived regularities of such a system are described by a differentiable function on an interval, the EC can be defined as the V-complexity of that function. EC is applied to the model of diffusion of cream in a cup of coffee. The perceived regularity of this model is given by the diffusion equation. The V-complexity of the solution of the equation starts at zero, quickly increases to a maximum and then decreases back to zero as the liquid reaches its equilibrium state. It is shown that this is also the result when a cellular automaton approach and the concept of Apparent Complexity is used.
arXiv:2510.05921v3 Announce Type: replace
Abstract: Large language models (LLMs) have achieved remarkable success in a wide range of natural language processing tasks and can be adapted through prompting. However, they remain suboptimal in multi-turn interactions, often relying on incorrect early assumptions and failing to track user goals over time, which makes such tasks particularly challenging. Prior works in dialogue systems have shown that long-term planning is essential for handling interactive tasks. In this work, we propose a prompt optimisation framework inspired by reinforcement learning, which enables such planning to take place by only modifying the task instruction prompt of the LLM-based agent. By generating turn-by-turn feedback and leveraging experience replay for prompt rewriting, our proposed method shows significant improvement in multi-turn tasks such as text-to-SQL and task-oriented dialogue. Moreover, it generalises across different LLM-based agents and can leverage diverse LLMs as meta-prompting agents. This warrants future research in reinforcement learning-inspired parameter-free optimisation methods.
arXiv:2605.17807v1 Announce Type: new
Abstract: Text-to-Image (T2I) generation has achieved remarkable progress in recent years. Meanwhile, reinforcement learning methods, particularly those based on Group Relative Policy Optimization (GRPO), have attracted widespread attention and been successfully applied to T2I tasks. However, the uniform sampling strategy commonly used during training often ignores the match between sample difficulty and the model's current learning capability, leading to low training efficiency. We argue that improving training efficiency requires continuously prioritizing prompts that match the model's evolving capability and remain actively learnable. To this end, we propose Curriculum Group Policy Optimization (CGPO), an adaptive curriculum training framework. During training, each prompt produces a group of images scored by a reward model. We use the variance of group rewards as an online proxy for prompt inconsistency. A higher variance suggests that the model has partially captured the prompt requirements but has not yet achieved stable mastery. Such prompts are more likely to provide useful learning signals, so we increase their sampling probabilities accordingly. Additionally, to address data imbalance in multi-category datasets, we design a category calibration method based on proportional fairness optimization, which balances training difficulty across categories. Experiments on GenEval, T2I-CompBench++, and DPG Bench demonstrate that our framework effectively improves generation performance.
arXiv:2605.18733v1 Announce Type: new
Abstract: Autoregressive video generation has improved rapidly in visual fidelity and interactivity, but it still suffers from long-term inconsistency and memory degradation. Most existing solutions either compress historical frames using predefined strategies or retrieve keyframes based on coarse implicit attention signals, both of which fail to handle evolving prompts with shifting entity references, leading to identity drift, character duplication, and attribute loss. To address this, we propose IAMFlow, a training-free identity-aware memory framework that explicitly models and tracks persistent entity identities, enabling consistent generation across prompt transitions. Specifically, an LLM extracts entities with visual attributes from each prompt and assigns unique global IDs for identity-aware memory, while a VLM asynchronously verifies and refines attributes from rendered frames, enabling explicit entity tracking in place of implicit similarity-based matching. To keep the proposed framework computationally practical, we design a systematic inference acceleration pipeline, including asynchronous visual verification, adaptive prompt transition, and model quantization, which achieves faster generation than existing baselines. Furthermore, we introduce NarraStream-Bench, a benchmark for narrative streaming video generation that features 324 multi-prompt scripts spanning six dimensions and a three-dimensional evaluation protocol that integrates both traditional metrics and multimodal large language model-based assessments. Extensive experiments show that IAMFlow, despite being training-free, achieves the best overall performance on NarraStream-Bench, outperforming the strongest baseline by 2.56 points, while achieving a 1.39$\times$ speedup over the most efficient baseline in the 60-second multi-prompt setting.
arXiv:2510.13068v4 Announce Type: replace
Abstract: Biosignals such as electroencephalography (EEG), electrocardiography (ECG), and electromyography (EMG) encode physiological activity across multiple temporal and spectral scales, yielding representations that are rich but challenging for machine learning. Foundation models trained to predict masked signal tokens have shown promise in learning generalizable biosignal representations, yet their performance depends on the tokenizer's ability to preserve high-frequency dynamics and reconstruct signals with high fidelity. We introduce NeuroRVQ, a modality-adaptive biosignal tokenizer family designed for high-fidelity signal reconstruction. To capture the full frequency spectrum, NeuroRVQ decomposes biosignals into frequency-specific representations via multi-scale temporal convolutions, each encoded into hierarchical RVQ codebooks to preserve high-frequency detail, combined with a novel phase-aware training loss that respects the circular topology of Fourier phase. By tuning the temporal resolution, number and size of temporal kernels and RVQ depth, this design adapts to the spectro-temporal characteristics of each biosignal modality. To validate that tokenizer quality drives downstream performance, we train a simple masked-token foundation model for each modality (NeuroRVQ-FM) using the corresponding NeuroRVQ tokenizer. The NeuroRVQ-FM family achieves competitive or superior downstream performance compared to existing modality-specific foundation models, demonstrating that high-fidelity tokenization is a critical factor for effective biosignal modeling.
arXiv:2510.14466v3 Announce Type: replace
Abstract: Large language models (LLMs) continue to struggle with low-resource languages, primarily due to limited training data, translation noise, and unstable cross-lingual alignment. To address these challenges, we propose LiRA (Linguistic Robust Anchoring for LLMs)-a plug-and-play framework that requires only lightweight fine-tuning on top of existing pretrained backbones. LiRA jointly optimizes representation stability and cross-lingual semantic consistency by combining two key components: Arca (Anchored Representation Composition Architecture), which aligns low-resource inputs to a shared English semantic space through anchor-based alignment and collaborative encoding; and LaSR (Language-coupled Semantic Reasoner), a lightweight, language-aware head that enforces consistency regularization for unified cross-lingual understanding, retrieval, and reasoning. We theoretically show that under controlled anchoring error and translation-induced bias, LiRA guarantees bounded representation deviation and stable downstream performance under local Lipschitz continuity. To facilitate research, we release a new multilingual product retrieval dataset covering five Southeast Asian and two South Asian languages. Extensive experiments across diverse low-resource benchmarks demonstrate consistent improvements in retrieval, ranking, question answering, and reasoning tasks. Code will be publicly available on GitHub, and the dataset will be hosted on Hugging Face.
arXiv:2503.13126v2 Announce Type: replace
Abstract: We study a variant of the Strang splitting for the time integration of the semilinear wave equation under the finite-energy condition on the torus $\mathbb{T}^3$. In the case of a cubic nonlinearity, we show almost second-order convergence in $L^2$ and almost first-order convergence in $H^1$. If the nonlinearity has a quartic form instead, we show an analogous convergence results, where the order is reduced by $1/2$ in both cases. To our knowledge these are the best convergence results available for the 3D cubic and quartic wave equations under the finite-energy condition. Our approach relies on continuous- and discrete-time Strichartz estimates. We also make use of the integration and summation by parts formulas to exploit cancellations in the error terms. Moreover, error bounds for a full discretization using the Fourier pseudo-spectral method in space are given. Finally, we discuss a numerical example indicating the sharpness of our theoretical results.
arXiv:2510.14699v2 Announce Type: replace
Abstract: Model checking for real-timed systems is a rich and diverse topic. Among the different logics considered, Metric Interval Temporal Logic (MITL) is a powerful and commonly used logic, which can succinctly encode many interesting timed properties especially when past and future modalities are used together. In this work, we develop a new approach for MITL model checking in the pointwise semantics, where our focus is on integrating past and maximizing determinism in the translated automata.
Towards this goal, we define synchronous networks of timed automata with shared variables and show that the past fragment of MITL can be translated in linear time to synchronous networks of deterministic timed automata. Moreover determinism can be preserved even when the logic is extended with future modalities at the top-level of the formula. We further extend this approach to the full MITL with past, translating it into networks of generalized timed automata (GTA) with future clocks (which extend timed automata and event clock automata). We present an SCC-based liveness algorithm to analyse GTA. We implement our translation in a prototype tool which handles both finite and infinite timed words and supports past modalities. Our experimental evaluation demonstrates that our approach significantly outperforms the state-of-the-art in MITL satisfiability checking in pointwise semantics on a benchmark suite of 72 formulas. Finally, we implement an end-to-end model checking algorithm for pointwise semantics and demonstrate its effectiveness on two well-known benchmarks.
arXiv:2510.15221v2 Announce Type: replace
Abstract: Affective computing has matured rapidly in laboratory settings, yet no prior dataset combines (i) months-to-years of duration, (ii) a naturalistic workplace context, (iii) a stable small-team social structure, and (iv) a fully passive sensing protocol that survives institutional review. We introduce WELD, the first dataset to satisfy all four. WELD comprises 733,780 per-frame seven-class facial-expression probability vectors from 49 employees of a Chinese software company over 30.1 months (Nov 2021 - May 2024) -- the longest naturalistic in-the-wild emotion corpus and the only multi-year corpus supporting both within-individual longitudinal and within-team relational analyses on the same subjects. Data are released under a four-tier access model with only aggregated probabilities publicly downloadable. We validate the corpus by replicating three established phenomena (+43.1% weekend valence boost; 13:00-trough diurnal cycle; Shanghai 2022 lockdown effect d=-0.40), and report four novel findings: (1) variance decomposition attributes 19.3% of daily-valence variance to between-person differences and 29.8% to month seasonality -- a quantitative ceiling for future predictive models; (2) Hidden Markov decomposition reveals six emotional regimes with asymmetric negative-state dwell times (16-18 d vs 3 d); (3) leave-one-person-out turnover prediction reaches AUC=0.79 yet a Cox concordance index of only 0.52, exposing a metric-trap when AUC is reported without survival-aware baselines; (4) the corpus reveals systematic over-prediction of "angry" by an off-the-shelf FER model on neutral Asian faces (0.194 vs ~0.05 Western priors), making WELD valuable for FER fairness audits. A complex-systems analysis of the corpus appears as a companion preprint (arXiv:2510.16046).
arXiv:2504.18302v3 Announce Type: replace
Abstract: We provide the first algorithm for computing an optimal tree decomposition for a given graph $G$ that runs in single exponential time in the feedback vertex number of $G$, that is, in time $2^{O(\text{fvn}(G))}\cdot n^{O(1)}$, where $\text{fvn}(G)$ is the feedback vertex number of $G$ and $n$ is the number of vertices of $G$. On a classification level, this improves the previously known results by Chapelle et al. [Discrete Applied Mathematics '17] and Fomin et al. [Algorithmica '18], who independently showed that an optimal tree decomposition can be computed in single exponential time in the vertex cover number of $G$.
One of the biggest open problems in the area of parameterized complexity is whether we can compute an optimal tree decomposition in single exponential time in the treewidth of the input graph. The currently best known algorithm by Korhonen and Lokshtanov [STOC '23] runs in $2^{O(\text{tw}(G)^2)}\cdot n^4$ time, where $\text{tw}(G)$ is the treewidth of $G$. Our algorithm improves upon this result on graphs $G$ where $\text{fvn}(G)\in o(\text{tw}(G)^2)$. On a different note, since $\text{fvn}(G)$ is an upper bound on $\text{tw}(G)$, our algorithm can also be seen either as an important step towards a positive resolution of the above-mentioned open problem, or, if its answer is negative, then a mark of the tractability border of single exponential time algorithms for the computation of treewidth.
arXiv:2510.15712v3 Announce Type: replace
Abstract: How hard is it to find a local optimum? If we are given a graph and want to find a locally maximal cut--meaning that the number of edges in the cut can't be improved by moving a single vertex from one side to the other--then just iterating improving steps finds a local maximum in $ |E|$ steps. If, on the other hand, the edges are weighted, this problem becomes hard for the class PLS (Polynomial Local Search). We are interested in optimization problems with lexicographic costs. For Max-Cut this would mean that the edges $e_1,\dots, e_m$ have costs $c(e_i) = 2^i$. For such a cost function finding a global Max-Cut is easy. In contrast, we show that it is PLS-complete to find an assignment for a 4-CNF formula that is locally maximal (when the clauses have lexicographic weights); and also for a 3-CNF when we allow switching two variables at a time. We use these results to answer a question in Scheder and Tantow, who showed that finding a lexicographic local minimum of a string $s \in \{0,1\}^n$ under the action of a list of given permutations $\pi_1, \dots, \pi_k \in S_{n}$ is PLS-complete. They ask whether the problem stays PLS-complete when the $\pi_1,\dots,\pi_k$ commute, i.e., generate an Abelian subgroup $G$ of $S_n$. We show that it does, and in fact stays PLS-complete even (1) when every element in $G$ has order two or (2) when $G$ is cyclic. Additionally, we use it to further investigate the complexity of computing pure $\alpha$-Nash equilibria in congestion games. Using lexicographic 4-SAT, we obtain a simple proof of the PLS-completeness originally shown by Skopalik and V\"ocking that can be extended to exponential and polynomial delay functions with positive coefficients. The number of strategies per player and players per resource is bounded. However, the degree of the polynomials is not bounded by a constant.
arXiv:2510.16252v2 Announce Type: replace
Abstract: Reinforcement learning (RL) for web agents demands environments that are both effective for evaluation and efficient enough for large-scale on-policy training. Current web environments fall short: server-side Docker setups are too resource-intensive for massive parallel rollouts, while browser-side interfaces produce noisy observations, execute actions unreliably under modern single-page applications, and omit visual interactivity cues.
We introduce WebServ, a full-stack, RL-ready web environment that addresses these limitations end-to-end. On the server side, WebServ uses Incus containers with block-level copy-on-write, reducing launch latency by ~5x and persistent storage by ~240x, enabling 200+ concurrent isolated environments on a single host. On the browser side, WebServ provides a compact, site-agnostic observation and action interface derived automatically from the DOM with human-aligned interactivity cues, and a robust action execution backend using network-aware waiting for reliable SPA support.
On WebArena-Lite, WebServ achieves state-of-the-art single-prompt results, with controlled comparisons confirming consistent gains across GPT-4o, OpenAI-o3, and Llama-3.1-8B over vanilla WebArena. We further train Qwen3-4B and Qwen3-30B-A3B with RL entirely within WebServ; the RL-trained 4B model achieves 55.5% mean accuracy, surpassing both Claude 4.5 Sonnet (50.0%) and the RL-trained 8B model from WebAgent-R1 (51.8%).
arXiv:2605.18693v1 Announce Type: new
Abstract: As LLM agents are increasingly built around reusable skills, a central challenge is no longer only whether agents can use provided skills, but whether they can generate correct, reusable, and executable skills from repositories and documents. Existing benchmarks primarily evaluate the efficacy of given skills or the ability of agents to solve downstream tasks from raw context, but they do not isolate skill generation itself as the object of study. We introduce SkillGenBench, a benchmark for evaluating skill generation pipelines under a unified and controlled protocol. In SkillGenBench, a generator receives raw corpora and produces standardized skill artifacts, which are then executed under fixed harnesses and assessed with unified evaluation procedures. The benchmark covers two generation regimes: task-conditioned generation, where a task-specific skill is synthesized after the task is revealed, and task-agnostic generation, where a reusable skill library must be distilled before downstream tasks are known. It also spans two complementary procedural sources: repository-grounded instances, where procedures are distributed across code, configuration, and scripts, and document-grounded instances, where procedures and constraints must be distilled from long-form text. We provide standardized task specifications, pinned environments, and evaluation protocols centered on deterministic execution-based checks, supplemented by auxiliary signals for diagnosis. Experiments across a range of skill-generation methods and backbones show substantial performance variation, highlight the difficulty of reusable skill distillation, and reveal distinct failure modes in skill generation from software repositories versus long-form documents. SkillGenBench establishes a reproducible testbed for studying skill generation as an independent research problem in agent systems.
arXiv:2604.15929v2 Announce Type: replace
Abstract: The goal of multilingual speech technology is to facilitate seamless communication between individuals speaking different languages, creating the experience as though everyone were a multilingual speaker. To create this experience, speech technology needs to address several challenges: Handling mixed multilingual input, specific vocabulary, and code-switching. However, there is currently no dataset benchmarking this situation. We propose a new benchmark to evaluate current Automatic Speech Recognition (ASR) systems, whether they are able to handle these challenges. The benchmark consists of bilingual discussions on scientific papers between multiple speakers, each conversing in a different language. We provide a standard evaluation framework, beyond Word Error Rate (WER) enabling consistent comparison of ASR performance across languages. Experimental results demonstrate that the proposed dataset is still an open challenge for state-of-the-art ASR systems. The dataset is available in https://huggingface.co/datasets/goodpiku/muscat-eval.
Keywords: multilingual, speech recognition, audio segmentation, speaker diarization
arXiv:2605.18680v1 Announce Type: new
Abstract: Metaverse platforms rely on creator-driven marketplaces where avatars are assembled from discrete, taxonomy-labeled 3D assets (e.g., tops, bottoms, shoes, accessories) under strict category and topology constraints. While users increasingly expect free-form text control, text-only retrieval is brittle: natural language is ambiguous with respect to platform taxonomies, metadata is often noisy or informal, and independently retrieved components can be stylistically inconsistent or geometrically incompatible. We propose \textbf{CMAG}, a concept-scaffolded retrieval and verified composition framework for marketplace avatar generation. Given a prompt, CMAG first synthesizes an intermediate 3D concept scaffold that disambiguates intent beyond text by providing global spatial and stylistic context. In parallel, a view-aware part discovery module extracts localized visual evidence via prompt decomposition and text-grounded segmentation. A prompt-conditioned taxonomy router enforces category coverage and resolves semantic-to-taxonomic mismatch, after which a hybrid category-wise retriever combines part-based fusion with a concept-residual fallback using feature suppression. Finally, an agentic vision--language model filters and re-ranks candidates across categories and drives an iterative verification loop to assemble prompt-faithful, topologically consistent avatars from catalog assets. We evaluate CMAG on diverse compositional prompts and demonstrate improved retrieval robustness and compositional correctness compared to strong baselines, highlighting the importance of 3D concept scaffolding under prompt ambiguity.
arXiv:2605.18653v1 Announce Type: new
Abstract: Micro-video popularity prediction (MVPP) forecasts the popularity a newly uploaded short-form video will attract within a fixed number of days after upload. This task supports downstream applications in recommendation, advertising, and creator analytics, yet the problem is hard since virality depends on external trends rather than video content alone. Prior MVPP methods incorporate context by retrieving similar videos from platform-internal corpora, however historical neighbors cannot reveal whether a topic is currently trending, controversial, or already saturated across the open web. To this end, we reformulate MVPP as open-web grounded prediction and introduce WEBSHORTS, the first micro-video dataset that couples 14K videos with real-time open-web context collected at upload time, alongside daily view counts tracked over 7 days. The context for each video is organized as a structured evidence-card that captures the external attention landscape along three complementary web-context dimensions. We further propose SHORTS-CAST, a framework that generates dimension-wise rationales from the evidence-card to guide popularity regression, then adapts at deployment by selectively updating the context-to-popularity mapping when delayed labels reveal genuine trend shifts. In our experiments, SHORTS-CAST consistently outperforms content-only, video corpus retrieval-augmented, and online adaptation baselines under both offline and delayed-label online protocols, confirming that structured web context and trend-aware adaptation are jointly necessary for popularity forecasting under realistic deployment constraints in fast-evolving short-form video ecosystems.
arXiv:2605.18462v1 Announce Type: new
Abstract: Named entity recognition (NER) has been one of the essential preliminary steps in modern NLP applications. This report focuses on implementing the NER task on finetuning two pretrained models: (i) an encoder-only model (BERT) with a simple classification head, and (ii) a sequence-to-sequence model (T5) with few-shot prompts. Under the original 7-class tag and 3-class simplified tag schemes, BERT is applied a weighted cross-entropy for training loss, and T5 is fine-tuned with two validation strategies. It also conducted an ablation study with different hyperparameters. Moreover, the related analysis provides valuable insights into common errors in BERT and the two models' performance. Based on a bunch of performance metrics, this report aims to compare the above two architectures and explore their abilities in the sequence labelling task, laying the groundwork for further practical use cases.
arXiv:2510.24883v2 Announce Type: replace
Abstract: The 60 mV$/$decade subthreshold limit at room temperature, coined as the Boltzmann tyranny, remains a fundamental obstacle to the continued down-scaling of conventional transistors. While several strategies have sought to overcome this constraint through non-thermal carrier injection, most rely on ferroelectric-based or otherwise material-specific mechanisms that require complex fabrication and stability control. Here, we develop a universal theoretical framework showing that intrinsic memory effects in nanometric field-effect transistors can naturally bypass this limit. Within the Landauer-B\"uttiker quantum transport formalism, we incorporate charge-trapping mechanisms that dynamically renormalize the conduction band edge. The resulting analytical expression for the subthreshold swing explicitly links memory dynamics to gate efficiency, revealing that a reduced carrier generation rate or enhanced trapping activity leads to sub-thermal switching, thus breaking the Boltzmann barrier. The model captures key experimental features and provides clear, generalizable design principles, establishing memory-assisted transistors as a robust pathway toward ultra-low-power and multifunctional electronic architectures.