Classical and neural NLP: translation, question answering, tokenization and the evaluation of language understanding.
arXiv:2505.16227v4 Announce Type: replace-cross Abstract: Personalizing jargon detection and explanation is essential for making technical documents accessible to readers with diverse disciplinary ba...
Lingshu is a medical‑specialized multimodal large language model that addresses key limitations of existing medical MLLMs, such as narrow knowledge coverage, hallucinations, and weak reasoning. The authors curate a comprehensive dataset combining medical imaging, texts, and general‑domain data, then train Lingshu in multiple stages to embed medical expertise and improve task performance. They also introduce MedEvalKit, a unified evaluation framework, and demonstrate that Lingshu outperforms current open‑source multimodal models on multimodal QA, text‑based QA, and medical report generation.
arXiv:2507.22920v2 Announce Type: replace Abstract: The rapid advancement of large language models (LLMs) has intensified the need for effective mechanisms to transform continuous multimodal data int...
Despite the success of neural models in natural language processing, their black-box nature limits interpretability and conceals the linguistic phenomena underlying their predictions. We present SLITE...
Cross-domain Named Entity Recognition (CD-NER) aims to transfer the rich knowledge in the source domain to the target domain. Recent studies adopting decomposition or generation paradigms have achieve...
Electroencephalography (EEG) provides a non-invasive window into dynamic brain activity, yet modeling long-horizon EEG sequences remains challenging due to their high temporal complexity, substantial...
The paper presents the first large‑scale benchmark for uncertainty quantification (UQ) calibration in long‑form scientific question answering, evaluating four UQ methods on 685,000 responses from up to 20 large language models across seven datasets. It shows that instruction tuning leads to token‑level probability polarization, undermining token‑level uncertainty signals, while reasoning model families differ in how they handle this effect. Only semantic consistency—consistency of the final answer—provides well‑calibrated outputs, demonstrating that semantic calibration remains robust in multi‑step, dependency‑rich reasoning.
The paper introduces an end‑to‑end framework for personalized audio‑driven facial motion that operates in real time without look‑ahead. It combines a causal multi‑resolution motion tokenizer, which captures both global temporal context and fine articulatory details, with a multi‑modal style retriever that pulls stylistic priors from arbitrary reference footage using ongoing audio and motion queries. This approach allows high‑fidelity, identity‑consistent animation from just a few casually recorded clips, outperforming existing methods in lip‑sync accuracy, identity consistency, and perceived realism while maintaining real‑time streaming constraints.
The paper introduces a multimodal reasoning framework for cross‑domain visual question answering in Printed Circuit Board Assembly (PCBA) inspection, converting diverse data sources into a unified instruction format and generating verified reasoning traces. It proposes Task‑Aware Group Relative Policy Optimization (GRPO) that uses semantic, distance‑aware, and format rewards to improve choice‑based and counting tasks beyond exact‑match supervision. During inference, the system applies semantic consistency correction, self‑consistency voting, and multi‑model arbitration, achieving an overall score of 83.24 on the PCBA Standard‑to‑Real Grand Challenge leaderboard.
The paper investigates gender bias in machine translation evaluation metrics using an occupation-balanced subset of GAMBIT+ across seven English‑source language pairs, including a new German extension. It finds that masculine translations tend to receive higher scores and that biases align with stereotypical gender representations, though the strength varies by evaluator and language. The study highlights that assessing bias requires multiple dimensions beyond a single aggregate measure.
The paper investigates when intermediate reasoning traces benefit machine translation by examining models, languages, domains, and datasets. It finds that the optimal reasoning language depends on the model, reasoning length has a non‑monotonic effect on quality, and traces display recurring functional patterns. Using Hierarchical Meta‑Summarization, the authors uncover a shared structure of understanding/planning, translating/drafting, and refining/verifying, while also noting domain‑specific variations, suggesting that reasoning should be tailored rather than uniformly applied.
AgentVidBench is a new multi‑hop video question‑answering benchmark designed to evaluate spatial, temporal, and causal reasoning in multimodal large language models (MLLMs). Unlike existing tests that focus on simple scene queries or global summaries, AgentVidBench includes step‑by‑step solution traces to assess whether agents gather the necessary evidence to justify their answers. Experiments with 12 MLLMs show limited single‑turn performance, but integrating these models into agentic workflows improves both accuracy and trajectory scores, establishing AgentVidBench as a comprehensive testbed for future research on agentic video understanding.
VidOmni-Bench is a new benchmark for fine‑grained video understanding that asks models to verify whether each event in dense video captions is supported by the video. It contains 500 videos covering five complexity types and durations from 4 seconds to 90 minutes, and uses human‑verified sentence‑level labels to create hard negatives. Experiments show that Video‑LLMs often hallucinate events, struggle to detect incorrect descriptions, and exhibit varying weaknesses depending on video complexity and duration.
arXiv:2609. 21362v1 Announce Type: new Abstract: Conventional tokenizers represent text as characters or statistically derived subwords, overlooking the internal phonological structure of syllables and often requiring large vocabularies.
BEAT-Net is a supervised biomimetic framework for ECG diagnosis that incorporates QRS-centered tokenization and a hierarchical architecture mirroring a cardiologist’s workflow. It processes heartbeat sequences through morphological, spatial, temporal, and transformer-based stages, achieving an AUC of 0.924 on large benchmarks while using only 0.7 million parameters. The model outperforms the 39.5‑million‑parameter HeartLang foundation model on morphological form classification and demonstrates superior cross‑dataset generalization with only 35% of the training data.
LiteMedCoT-VL is a parameter‑efficient pipeline that transfers chain‑of‑thought reasoning from a 235B teacher model to a 2B student model using LoRA fine‑tuning on explanation‑enriched data. The approach enables a compact vision‑language model to perform medical visual question answering without relying on image captions, achieving 64.9% accuracy on the PMC‑VQA benchmark—an 11‑point improvement over the zero‑shot Qwen3‑VL‑4B baseline. Visual grounding analysis confirms that the model bases its predictions on image content rather than textual priors.
The paper introduces a training strategy for cascaded simultaneous speech translation that allows the system to dynamically decide how much of the source prefix to translate. By fine‑tuning a large language model (Qwen3‑8B) on stable prefixes—pairs of source prefixes and the longest shared translation with the full sentence—the authors enable contextual read‑write decisions beyond fixed wait‑k or target‑suffix deletion. Experiments on English‑to‑German, Japanese, and Chinese demonstrate that stable prefixes improve the quality‑latency tradeoff across various test sets.
VISPATH is a visual‑intent‑guided path reasoning framework designed for multimodal knowledge graph question answering (MM‑KGQA). It first identifies a reliable starting entity by fusing multimodal grounding with graph‑structural cues, then iteratively discovers and refines reasoning paths using hop‑specific multimodal intent and a reasoning‑chain pruning step. The framework is evaluated on the newly introduced VISPATH‑Bench, which tests two‑to‑four‑hop reasoning, and demonstrates consistent improvements over strong baselines, even surpassing GPT‑5.4 when using GPT‑4o as the backbone.
PhysioBench is a unified benchmark for physiological signal question answering that consolidates annotations from 22 public datasets into 61.4 million question‑answer pairs spanning 30 tasks. Each pair is linked to a specific signal segment and traceable to its source annotation. The benchmark evaluates 21 models—including large language models, vision‑language models, time‑series language models, and physiological signal foundation models—under three settings, revealing that no model consistently excels across all modalities and tasks, and that performance is sensitive to question wording.
arXiv:2609.21940v1 Announce Type: new Abstract: Long-term memory is essential for large language model (LLM) agents to maintain consistency and personalization over extended interactions. Existing me...