Existing methods like ensemble still struggle to aggregate these models with the desired higher-order communications. In this work, we propose Cola, a novel paradigm that coordinates multiple VLMs for visual reasoning.

What's happening in Myanmar's civil war as military holds elections? Voters in parts of Myanmar are heading to the polls on Sunday for an election that critics view as a bid by the country's generals to legitimise military rule, nearly five years after they overthrew the government of Nobel Laureate Aung San Suu Kyi. The multi-phased election is unfolding amid a raging civil war, with ethnic armed groups and opposition militias fighting the military for control of vast stretches of territory, stretching from the borderlands with Bangladesh and India in the west, across the central plains, to the frontiers with China and Thailand in the north and east. Another third will be covered during a second and third phase in January, while voting has been cancelled altogether in the remainder. Fighting, including air raids and arson, has intensified in several areas.

Abstract--This paper investigates sequence-to-sequence T ransformer models for automatic speech recognition (ASR) error correction in low-resource Burmese, focusing on different feature integration strategies including IP A and alignment information. T o our knowledge, this is the first study addressing ASR error correction specifically for Burmese. W e evaluate five ASR backbones and show that our ASR Error Correction (AEC) approaches consistently improve word-and character-level accuracy over baseline outputs. The proposed AEC model, combining IP A and alignment features, reduced the average WER of ASR models from 51.56 to 39.82 before augmentation (and 51.56 to 43.59 after augmentation) and improving chrF++ scores from 0.5864 to 0.627, demonstrating consistent gains over the baseline ASR outputs without AEC. Our results highlight the robustness of AEC and the importance of feature design for improving ASR outputs in low-resource settings.

In low-resource languages like Burmese, classification tasks often fine-tune only the final classification layer, keeping pre-trained encoder weights frozen. While Multi-Layer Perceptrons (MLPs) are commonly used, their fixed non-linearity can limit expressiveness and increase computational cost. This work explores Kolmogorov-Arnold Networks (KANs) as alternative classification heads, evaluating Fourier-based FourierKAN, Spline-based EfficientKAN, and Grid-based FasterKAN-across diverse embeddings including TF-IDF, fastText, and multilingual transformers (mBERT, Distil-mBERT). Experimental results show that KAN-based heads are competitive with or superior to MLPs. EfficientKAN with fastText achieved the highest F1-score (0.928), while FasterKAN offered the best trade-off between speed and accuracy. On transformer embeddings, EfficientKAN matched or slightly outperformed MLPs with mBERT (0.917 F1). These findings highlight KANs as expressive, efficient alternatives to MLPs for low-resource language classification.

Large Language Models (LLMs) equipped with modern Retrieval-Augmented Generation (RAG) systems often employ multi-turn interaction pipelines to interface with search engines for complex reasoning tasks. However, such multi-turn interactions inevitably produce long intermediate contexts, as context length grows exponentially with exploration depth. This leads to a well-known limitation of LLMs: their difficulty in effectively leveraging information from long contexts. This problem is further amplified in RAG systems that depend on in-context learning, where few-shot demonstrations must also be included in the prompt, compounding the context-length bottleneck. To address these challenges, we propose Mujica-MyGo, a unified framework for efficient multi-turn reasoning in RAG. Inspired by the divide-and-conquer principle, we introduce Mujica (Multi-hop Joint Intelligence for Complex Question Answering), a multi-agent RAG workflow that decomposes multi-turn interactions into cooperative sub-interactions, thereby mitigating long-context issues. To eliminate the dependency on in-context learning, we further develop MyGO (Minimalist Policy Gradient Optimization), a lightweight and efficient reinforcement learning algorithm that enables effective post-training of LLMs within complex RAG pipelines. We provide theoretical guarantees for MyGO's convergence to the optimal policy. Empirical evaluations across diverse question-answering benchmarks, covering both text corpora and knowledge graphs, show that Mujica-MyGO achieves superior performance.

Named entities are fundamental building blocks of knowledge in text, grounding factual information and structuring relationships within language. Despite their importance, it remains unclear how Large Language Models (LLMs) internally represent entities. Prior research has primarily examined explicit relationships, but little is known about entity representations themselves. We introduce entity mention reconstruction as a novel framework for studying how LLMs encode and manipulate entities. We investigate whether entity mentions can be generated from internal representations, how multi-token entities are encoded beyond last-token embeddings, and whether these representations capture relational knowledge. Our proposed method, leveraging _task vectors_, allows to consistently generate multi-token mentions from various entity representations derived from the LLMs hidden states. We thus introduce the _Entity Lens_, extending the _logit-lens_ to predict multi-token mentions. Our results bring new evidence that LLMs develop entity-specific mechanisms to represent and manipulate any multi-token entities, including those unseen during training. Our code is avalable at https://github.com/VictorMorand/EntityRepresentations .

Existing methods like ensemble still struggle to aggregate these models with the desired higher-order communications. In this work, we propose Cola, a novel paradigm that coordinates multiple VLMs for visual reasoning.

Despite growing research on bias in large language models (LLMs), most work has focused on gender and race, with little attention to religious identity. This paper explores how religion is internally represented in LLMs and how it intersects with concepts of violence and geography. Using mechanistic interpretability and Sparse Autoencoders (SAEs) via the Neuronpedia API, we analyze latent feature activations across five models. We measure overlap between religion- and violence-related prompts and probe semantic patterns in activation contexts. While all five religions show comparable internal cohesion, Islam is more frequently linked to features associated with violent language. In contrast, geographic associations largely reflect real-world religious demographics, revealing how models embed both factual distributions and cultural stereotypes. These findings highlight the value of structural analysis in auditing not just outputs but also internal representations that shape model behavior.

Large language models (LLMs) often struggle to perform multi-target reasoning in long-context scenarios where relevant information is scattered across extensive documents. To address this challenge, we introduce NeuroSymbolic Augmented Reasoning (NSAR), which combines the benefits of neural and symbolic reasoning during inference. NSAR explicitly extracts symbolic facts from text and generates executable Python code to handle complex reasoning steps. Through extensive experiments across seven languages and diverse context lengths, we demonstrate that NSAR significantly outperforms both a vanilla RAG baseline and advanced prompting strategies in accurately identifying and synthesizing multiple pieces of information. Our results highlight the effectiveness of combining explicit symbolic operations with neural inference for robust, interpretable, and scalable reasoning in multilingual settings.

The proliferation of generative AI and deceptive synthetic media threatens the global information ecosystem, especially across the Global Majority. This report from WITNESS highlights the limitations of current AI detection tools, which often underperform in real-world scenarios due to challenges related to explainability, fairness, accessibility, and contextual relevance. In response, WITNESS introduces the Truly Innovative and Effective AI Detection (TRIED) Benchmark, a new framework for evaluating detection tools based on their real-world impact and capacity for innovation. Drawing on frontline experiences, deceptive AI cases, and global consultations, the report outlines how detection tools must evolve to become truly innovative and relevant by meeting diverse linguistic, cultural, and technological contexts. It offers practical guidance for developers, policy actors, and standards bodies to design accountable, transparent, and user-centered detection solutions, and incorporate sociotechnical considerations into future AI standards, procedures and evaluation frameworks. By adopting the TRIED Benchmark, stakeholders can drive innovation, safeguard public trust, strengthen AI literacy, and contribute to a more resilient global information credibility.