Researchers evaluated four state-of-the-art zero-shot cross-lingual voice cloning systems across six languages, finding that Arabic remains particularly challenging. The study, presented at IWSLT 2026…
Researchers presented AURA-ST, a three-stage modular pipeline for low-resource speech-to-text translation at IWSLT 2026. The system bypasses traditional cross-attention by treating projected acoustic …
Researchers introduced Patent-CR, the first English dataset for patent claim revision, containing rejected and granted patent claims. Evaluating large language models, they found GPT-4 performed best …
Researchers at the 2025 Conference on Empirical Methods in Natural Language Processing introduced PatentScore, a multi-dimensional evaluation framework for LLM-generated patent claims. PatentScore int…
Researchers at the 64th Annual Meeting of the Association for Computational Linguistics introduced a dual-task paradigm combining arithmetic and sentence comprehension to study how language models all…
Researchers at the Association for Computational Linguistics introduced a Dual-Phase Self-Evolution (DPSE) framework for large language models that jointly optimizes user preference adaptation and dom…
Researchers at French institutions found that diversity-driven sampling can reduce pre-training dataset size by up to 94% and training time by 73% while maintaining performance in ModernBERT models. I…
Researchers Bhavik Chandna and Procheta Sen introduced a counterfactual explanation framework for retrieval models that identifies which terms need to be added to a document to improve its ranking, ad…
Researchers at ACL 2026 introduced a method to reuse LLM hidden states for classification tasks, training lightweight probes on token- and layer-selective representations to eliminate separate guard m…
Researchers at the Association for Computational Linguistics found that large language models' numerical accuracy drops significantly when numbers are presented in underrepresented scripts or formats,…
Researchers introduced TRuST, a large-scale dataset with ~300k annotations for toxicity recognition, unifying prior resources through a synthesized definition. Benchmarking showed fine-tuned PLMs outp…
Researchers introduced CoLabScience, a proactive AI assistant that uses a reinforcement learning framework called PULI to autonomously intervene in biomedical research discussions. The system, trained…
Researchers from multiple institutions published a comprehensive survey on Process Reward Models (PRMs) at ACL 2026, covering data generation, model construction, and usage for step-level reasoning ev…
Researchers introduced DySECT, a dynamic self-evolving extraction system that uses a closed-loop cycle where an LLM extracts triples to populate a knowledge base, which then improves the LLM via promp…
Drexel University researchers analyzed over 4 million Reddit posts across 47 mental-health subreddits and found that users who interact with AI chatbots for task-oriented support, such as planning or …