Natural Language Processing news and analysis on Web Pulse: 5475 curated articles tracking the latest Natural Language Processing developments, tools, and research, updated continuously from vetted sources.
Corti, a healthcare AI company, provides clinical-grade APIs for speech-to-text, ICD-10 and CPT medical coding, and note generation, along with an agentic framework of over 20 pre-built healthcare agents, targeting EHR v…
Hugging Face user @furkanyllmz's authorship claim on the paper 'MoganColBERT-TR: A Late-Interaction Multi-Vector Retrieval Model for Turkish' (arXiv 2608.26344) has failed approval for the second time, prompting a suppor…
Paxa Labs, an AI research lab in Bangkok, has launched a suite of Thai-language AI models delivered through a single API, including text-to-speech, OCR, and translation. The models are designed to handle the unique chall…
OpenAI has introduced a new Codex skill called 'write-persuasively' that guides AI models to draft, rewrite, and critique persuasive text by first interviewing the user to identify the purpose, audience, desired action, …
A developer built Trainlog, a voice-first AI sports journal that uses speech recognition and large language models to help users log workouts without friction. The project uses a deliberately simple stack—React 19, Fireb…
Moonraker Inspector, an AI construction inspection system, sends one email per trade with only that trade's findings, each defect carrying a clause reference and a named supervisor who verified it, after a contractor's '…
Google launched Gemini 3.5 Transcribe on August 26, replacing Chirp 3 as its speech-to-text model and making it the first STT model sold natively through the Gemini API, with two variants in public preview: gemini-3.5-tr…
Famulor added a Data migration panel under Settings → Workspace that lets Owners and Admins import Retell, Vapi, or Synthflow assistants as inactive copies, with no plan gate. The importer copies only the assistant name,…
A developer has developed a cross-modal knowledge distillation approach to aid in the revitalization of heritage languages such as Ainu, which have limited text data but richer audio and visual archives. The method uses …
Researchers propose Relational Over-Regularization (ROR), a sentence-pair-level signal showing LLMs produce inter-sentence transition variance that deviates from human writing, and validate it across four benchmarks (p <…
Researchers proposed a training-time explainability framework that aligns model reasoning with human-annotated rationales, improving both classification performance and interpretability for multilingual hate speech detec…
A new audit framework, the Cross-Platform Fairness Evaluation (CPFE), applied to four transformer models (BERT, RoBERTa, Emotion-DistilRoBERTa, GoEmotions-RoBERTa) trained on a Kaggle mental health corpus (n=35,556) foun…
Hugging Face and Voice Arena have added the first Global South language to the Open ASR Leaderboard, introducing Hindi and Indian English evaluation sets called Monsoon en-IN and Monsoon hi-IN, which include 4,888 speake…
OpenAI has added a native meetings plugin to ChatGPT that records audio, generates transcripts, and creates AI-generated notes, available to Team, Enterprise, and Education subscribers on macOS since June 2025 and to Plu…
Tim O'Reilly discusses the integration of AI into the writing process, sharing his experiences and insights on how AI tools can assist authors. He explores the potential benefits and challenges of AI-assisted writing, em…
Google released Gemini 3.5 Transcribe, a speech-to-text model supporting real-time transcription across 85+ languages with custom vocabulary recognition, and Gemini Omni 1.1 Flash, a video model with scene extension, fra…
Google has added agentic productivity features to Gemini Live, enabling voice-based access to Daily Brief, email triage, and help from Spark. The update positions Gemini Live as a tool for everyday work tasks, though Goo…
Tavus released Sparrow-2, a conversational AI model that processes the full audio stream—including breaths, sighs, and background noise—rather than isolating a single speaker, aiming to improve turn-taking and environmen…
Telnyx has released a new open-source example, 'shipment-agent,' that turns a package shipment into a durable AI agent capable of proactive SMS updates, understanding customer replies via Telnyx AI Inference, and answeri…
Tether's AI team has developed dedicated, resource-optimized translation models that run entirely on-device, requiring only 21-35 MB per language pair and translating at ~46ms per sentence, which is ~78 times faster than…