Majority voting has proven effective for close-ended question answering by aggregating parallel reasoning traces. However, it is not directly applicable to open-ended reasoning, such as code generation and web-based deep research, where a "majority" over complete solutions is ill-defined. Empirically, it matches or surpasses majority voting on AIME and GPQA, while delivering consistent gains on open-ended coding tasks: on LiveCodeBench (hard), pass@1 improves by +8.28% for DeepCoder-14B-Preview and +7.58% for Qwen3-8B. These results demonstrate that parallel test-time scaling can benefit open-ended reasoning without relying on voting over complete outputs. Recent advances in Large Language Models (LLMs) have been driven by test-time compute scaling. As evidenced by OpenAI's o1 (OpenAI, 2024), DeepSeek-R1 (Guo et al., 2025), etc., models generate extended "think" segments that reflect intermediate hypotheses, derivations, and self-corrections prior to emitting the final answer (Chen et al., 2025b; Y ang et al., 2025c). Such sequential test-time scaling has established a new paradigm: increasing the inference-time computation (e.g., longer reasoning traces) often leads to improved accuracy and problem-solving capability. Y et simply lengthening the chain has diminishing returns and can even hurt, e.g., overthinking (Chen et al., 2024; Cuadron et al., 2025), with studies showing that correct answers often appear in shorter traces (Zeng et al., 2025).

Hate speech recognition in low-resource languages remains a difficult problem due to insufficient datasets, orthographic heterogeneity, and linguistic variety. Bangla is spoken by more than 230 million people of Bangladesh and India (West Bengal). Despite the growing need for automated moderation on social media platforms, Bangla is significantly under-represented in computational resources. In this work, we study Subtask 1A and Subtask 1B of the BLP 2025 Shared Task on hate speech detection. We reproduce the official baselines (e.g., Majority, Random, Support Vector Machine) and also produce and consider Logistic Regression, Random Forest, and Decision Tree as baseline methods. We also utilized transformer-based models such as DistilBERT, BanglaBERT, m-BERT, and XLM-RoBERTa for hate speech classification. All the transformer-based models outperformed baseline methods for the subtasks, except for DistilBERT. Among the transformer-based models, BanglaBERT produces the best performance for both subtasks. Despite being smaller in size, BanglaBERT outperforms both m-BERT and XLM-RoBERTa, which suggests language-specific pre-training is very important. Our results highlight the potential and need for pre-trained language models for the low-resource Bangla language.

Autonomous driving technology is spearheading a transformation in the global automotive industries, and its safe and reliable implementation is the core prerequisite for large-scale adoption (Ren et al., 2025). Comprehensive testing and evaluation of autonomous driving systems (ADSs) are essential to ensuring their safety, in which the identification and generation of safety-critical scenarios represent a core challenge (Yang et al., 2025). "Safety-critical scenarios" specifically refer to rare driving situations with potentially high risks (Ding et al., 2023). Conducting tests under such scenarios enables effective evaluation of the ADSs' safety performance, as well as the clarification and iterative refinement of its Operational Design Domain (ODD). However, due to the rarity of safety-critical scenarios in naturalistic driving environments (Feng et al., 2023), real-world road testing is inefficient and cost-prohibitive, making it unsuitable for large-scale testing of high-level ADSs. As a more efficient and practical solution, simulation-based testing has garnered significant industrial and scholarly attention (Sun et al., 2022). In recent years, engineers in enterprises generally extract safety-critical testing scenarios by directly replaying vehicle-collected data in simulation environments (Liu et al., 2024), while some researchers achieve accelerated sampling of safety-critical scenarios through optimization-based search within a predefined scenario parameter space (Wu et al., 2024, 2026). However, the background vehicles (BVs) in the safety-critical testing scenarios generated by the aforementioned methods exhibit fixed behaviors and cannot dynamically respond to the actions of the vehicle under test (VUT). As a remedy, some other studies have introduced reinforcement learning to train adversarial BV driver models, thereby constructing naturalistic adversarial driving environments (NADE) (Feng et al., 2021) or evolving scenarios (Ma et al., 2024; Wu et al., 2025).

Reasoning-centric video object segmentation is an inherently complex task: the query often refers to dynamics, causality, and temporal interactions, rather than static appearances. Yet existing solutions generally collapse these factors into simplified reasoning with latent embeddings, rendering the reasoning chain opaque and essentially intractable. We therefore adopt an explicit decomposition perspective and introduce ReVSeg, which executes reasoning as sequential decisions in the native interface of pretrained vision language models (VLMs). Rather than folding all reasoning into a single-step prediction, ReVSeg executes three explicit operations -- semantics interpretation, temporal evidence selection, and spatial grounding -- aligning pretrained capabilities. We further employ reinforcement learning to optimize the multi-step reasoning chain, enabling the model to self-refine its decision quality from outcome-driven signals. Experimental results demonstrate that ReVSeg attains state-of-the-art performances on standard video object segmentation benchmarks and yields interpretable reasoning trajectories. Project page is available at https://clementine24.github.io/ReVSeg/ .

The globalization of education and rapid growth of online learning have made localizing educational content a critical challenge. Lecture materials are inherently multimodal, combining spoken audio with visual slides, which requires systems capable of processing multiple input modalities. To provide an accessible and complete learning experience, translations must preserve all modalities: text for reading, slides for visual understanding, and speech for auditory learning. We present \textbf{BOOM}, a multimodal multilingual lecture companion that jointly translates lecture audio and slides to produce synchronized outputs across three modalities: translated text, localized slides with preserved visual elements, and synthesized speech. This end-to-end approach enables students to access lectures in their native language while aiming to preserve the original content in its entirety. Our experiments demonstrate that slide-aware transcripts also yield cascading benefits for downstream tasks such as summarization and question answering. We release our Slide Translation code at https://github.com/saikoneru/image-translator and integrate it in Lecture Translator at https://gitlab.kit.edu/kit/isl-ai4lt/lt-middleware/ltpipeline}\footnote{All released code and models are licensed under the MIT License.

The emergence of Large Language Models (LLMs) within the Traditional Chinese Medicine (TCM) domain presents an urgent need to assess their clinical application capabilities. However, such evaluations are challenged by the individualized, holistic, and diverse nature of TCM's "Syndrome Differentiation and Treatment" (SDT). Existing benchmarks are confined to knowledge-based question-answering or the accuracy of syndrome differentiation, often neglecting assessment of treatment decision-making. Here, we propose a comprehensive, clinical case-based benchmark spearheaded by TCM experts, and a specialized reward model employed to quantify prescription-syndrome congruence. Data annotation follows a rigorous pipeline. This benchmark, designated TCM-BEST4SDT, encompasses four tasks, including TCM Basic Knowledge, Medical Ethics, LLM Content Safety, and SDT. The evaluation framework integrates three mechanisms, namely selected-response evaluation, judge model evaluation, and reward model evaluation. The effectiveness of TCM-BEST4SDT was corroborated through experiments on 15 mainstream LLMs, spanning both general and TCM domains. To foster the development of intelligent TCM research, TCM-BEST4SDT is now publicly available.

Radiology reporting is an essential yet time-consuming and error-prone task for radiologists in clinical examinations, especially for volumetric medical images. Rigorous quality control is also critical but tedious, ensuring that the final report meets clinical standards. Existing automated approaches, including radiology report generation methods and medical vision-language models, focus mainly on the report generation phase and neglect the crucial quality control procedure, limiting their capability to provide comprehensive support to radiologists. We propose Radiologist Copilot, an agentic AI assistant equipped with orchestrated tools designed for automated radiology reporting with quality control. Leveraging large language models as the reasoning backbone, the agentic system autonomously selects tools, plans, and executes actions, emulating the behavior of radiologists throughout the holistic radiology reporting process. The orchestrated tools include region localization, think with image paradigm directed region analysis planning, strategic template selection for report generation, quality assessment and feedback-driven adaptive refinement for quality control. Therefore, Radiologist Copilot facilitates accurate, complete, and efficient radiology reporting, assisting radiologists and improving clinical efficiency. Experimental results demonstrate that Radiologist Copilot significantly surpasses other state-of-the-art methods in radiology reporting. The source code will be released upon acceptance.

Aligning Large Language Models (LLMs) with human preferences typically relies on external supervision, which faces critical limitations: human annotations are scarce and subjective, reward models are vulnerable to reward hacking, and self-evaluation methods suffer from prompt sensitivity and biases. In this work, we propose stable rank, an intrinsic, annotation-free quality signal derived from model representations. Stable rank measures the effective dimensionality of hidden states by computing the ratio of total variance to dominant-direction variance, capturing quality through how information distributes across representation dimensions. Empirically, stable rank achieves 84.04% accuracy on RewardBench and improves task accuracy by an average of 11.3 percentage points over greedy decoding via Best-of-N sampling. Leveraging this insight, we introduce Stable Rank Group Relative Policy Optimization (SR-GRPO), which uses stable rank as a reward signal for reinforcement learning. Without external supervision, SR-GRPO improves Qwen2.5-1.5B-Instruct by 10% on STEM and 19% on mathematical reasoning, outperforming both learned reward models and self-evaluation baselines. Our findings demonstrate that quality signals can be extracted from internal model geometry, offering a path toward scalable alignment without external supervision.

Low-resource African languages remain underrepresented in sentiment analysis research, resulting in limited lexical resources and reduced model performance in multilingual applications. This gap restricts equitable access to Natural Language Processing (NLP) technologies and hinders downstream tasks such as public-health monitoring, digital governance, and financial inclusion. To address this challenge, this paper introduces TriLex, a three-stage retrieval-augmented framework that integrates corpus-based extraction, cross-lingual mapping, and Retrieval-Augmented Generation (RAG) driven lexicon refinement for scalable sentiment lexicon expansion in low-resource languages. Using an expanded lexicon, we evaluate two leading African language models (AfroXLMR and AfriBERTa) across multiple case studies. Results show that AfroXLMR consistently achieves the strongest performance, with F1-scores exceeding 80% for isiXhosa and isiZulu, aligning with previously reported ranges (71-75%), and demonstrating high multilingual stability with narrow confidence intervals. AfriBERTa, despite lacking pre-training on the target languages, attains moderate but reliable F1-scores around 64%, confirming its effectiveness under constrained computational settings. Comparative analysis shows that both models outperform traditional machine learning baselines, while ensemble evaluation combining AfroXLMR variants indicates complementary improvements in precision and overall stability. These findings confirm that the TriLex framework, together with AfroXLMR and AfriBERTa, provides a robust and scalable approach for sentiment lexicon development and multilingual sentiment analysis in low-resource South African languages.

ABSTRACT Dialogue-Based Generalized Referring Expressions Comprehension (GREC) requires models to ground the expression and unlimited targets in complex visual scenes while resolving coreference across a long dialogue context. However, existing systems struggle under distribution shift between training and evaluation domains, a gap exacerbated by the scarcity of annotated dialogue grounding data. We address this challenge with a three-tier data-synthesis method that balances realism and controllability to produce scalable supervision for dialogue-conditioned grounding. Fine-tuning on the synthesized data yields consistent, substantial improvements over prior approaches across standard evaluation metrics. Index T erms-- Visual Grounding, Referring Expression Comprehension, Generalized Referring Expression Comprehension, Coreference, Data Synthesis 1. INTRODUCTION Referring Expression Comprehension (REC) - the task of locating a target referred to by a natural language description.