Natural Language Processing news and analysis on Web Pulse: 5440 curated articles tracking the latest Natural Language Processing developments, tools, and research, updated continuously from vetted sources.
A frontend-backend architecture that lets full-duplex speech-to-speech models delegate tool calls to a text-based backend LLM achieved 92-97% tool-call recall and 81.2% accuracy in rejecting irrelevant calls in a single-…
The YNU-HPCC team placed in subtask A of SemEval-2025 Task 11 with a RoBERTa-based system that scored an official ranking score of 0.44 across all languages, according to the team's arXiv paper. The team modified the mod…
A new arXiv paper (2609.20068v1) unifies marginal utility, matrix factorization, and the transformer Key-Value cache under a single allocation rule that retains the top dimensions whose eigenvalue exceeds the shadow pric…
VisKG-LM, a new method detailed in arXiv paper 2609.19158v1, compiles retrieved knowledge graphs once offline into cached images that a language model reads as read-only visual memory, decoupling graph encoding from lang…
A training-free method that repeatedly samples completions of a single prompt at elevated temperature and applies Principal Component Analysis to pooled hidden activations can discover prompt-conditional stylistic axes i…
Researchers constructed UC-Bench, a human-annotated benchmark for detecting user-side implicit conflicts in Human-LLM dialogue, and found existing LLMs struggle with the task, especially when conflicts stem from implicit…
A new benchmark called Neo-Classic, introduced in arXiv paper 2609.19154v1, shows that state-of-the-art large language models lose 20 to 50 percent accuracy when moving from historical Classical Chinese Poetry to strictl…
Researchers introduced QVAC Genesis III, a 191.43B-token STEM-focused synthetic corpus spanning 19 domains, built via a dual generation strategy that uses a weak edge-scale student model's failures as corrective explanat…
A new arXiv paper (2609.19465v1) from unnamed researchers proposes a dependency-graph framework that formalizes compositional reasoning in language models into three levels of increasing complexity, and reports a consist…
A new arXiv paper (2609.20152v1) introduces MTVA-Bench, a benchmark that evaluates the language model inside cascaded voice agents under the same conditions it faces in production, covering 49 agents, 490 reviewed scenar…
The Modality Discrepancy Transformer (MDT), a conflict-aware multimodal fusion model built on the framework of Bekhouche et al., achieved 0.7408 Macro F1 on the labelled test split and 0.7368 on the private leaderboard o…
TrioRAG, a graph-free multimodal retrieval-augmented generation framework, matches or outperforms graph-based systems across three benchmarks while cutting total cost and speeding per-query inference by 1.6 to 2.3 times,…
A developer has published a from-scratch Python walkthrough for building a retrieval-augmented generation (RAG) system, covering chunking, embeddings, a minimal vector store, hybrid retrieval, re-ranking, and cited gener…
A developer quantized a fine-tuned DistilBERT Banking77 support-ticket classifier to ONNX int8, shrinking the model from 256 MB to 64 MB so it could run entirely client-side in the browser via Transformers.js with no bac…
SAP CEO Christian Klein told Fortune that "the end of the keyboard is near," predicting voice control and AI will replace much of the manual data entry in SAP's own systems within two to three years. Klein said the under…
Alibaba's Qwen3-Omni Flash, a native omni model handling text, image, audio, and video in a single system with a reported 1 million token context window, led several audio-visual understanding benchmarks shared by Qwen, …
A production voice AI system serving residential property-management companies now scores every completed call with a separate evaluator model, internally called CSAT AI, rather than sampling calls for human review. The …
Appen published an analysis warning that OpenAI's Whisper and medical speech APIs are producing transcription errors that can be potentially fatal in clinical settings. Appen cites its own 30 years of AI data expertise, …
A developer building a real-time Persian voice conversation pipeline for a Reachy Mini robot is seeking advice on running both the LLM and text-to-speech locally on a GTX 1650 Ti with 4GB of VRAM, after cloud inference v…
A developer built a production voice agent for travel booking inside an existing Amazon Connect environment, reaching a 92% successful end-to-end transaction rate and saving 10 to 15 minutes per automated booking. The sy…