Natural Language Processing news and analysis on Web Pulse: 5575 curated articles tracking the latest Natural Language Processing developments, tools, and research, updated continuously from vetted sources.
Researchers introduced destroR, a pipeline with three attack strategies—paraphrase, back-translation, and one-hot word-swap—that revealed Bangla sentiment classifiers are vulnerable to adversarial attacks. Standard attac…
China's NLP community is preparing for the first ChineseBabyLM challenge in 2026 at the NLPCC conference, tasking researchers with building language models using 100 million Chinese tokens across three tracks: Natural La…
Researchers are shifting AI video detection from pixel-level artifact analysis to high-level factual fidelity verification, driven by the increasing realism of AI-generated content (AIGC-V). A new taxonomy called Vision-…
Datarelax Universe launched an agentic data platform that generates dashboards and models in minutes, featuring an integrated AI assistant capable of producing DBML schemas from natural language, modifying existing model…
Piotr has released StageWhisper Lite, a free, on-device macOS app for meeting transcription, summaries, and action items using local models like parakeet and gemma 4 or BYO AI via OpenClaw or Hermes Agent. The app requir…
A new article from Wonder Tools outlines five tactics for using AI to sharpen thinking rather than outsourcing it, including using AI as an intellectual sparring partner and employing voice AI tools for bionic dictation.…
Revenue Growth Agent expanded from AI meeting preparation into a broader sales execution system that carries buyer context through discovery, HubSpot workflows, proposals and Statements of Work, the Cape Coral, Florida c…
Meta made its Business AI agent available worldwide on June 3, bundling it into WhatsApp's premium business tiers and charging large companies by the token. The agent, which answers questions, recommends products, and bo…
Turnitin Report, a free independent AI checker for student papers, lets users paste essays to receive an estimated AI score, paragraph-level signals, and writing review notes before submission. The tool analyzes patterns…
A developer advises that production AI voice agents should start speaking within 800ms-1.2 seconds of a caller finishing a sentence, as delays past 2 seconds cause callers to assume the line dropped or hang up. The devel…
A developer has built a searchable chat interface for Stanford's CS229 machine learning course notes, allowing users to explore the material with citations back to the original source. The tool, hosted at chat.pageindex.…
A developer improved the perceived speed of a local voice assistant by streaming sentences instead of audio, using sentence-level pipelining and overlapping synthesis with playback. The fix involved a regex, a queue, and…
Large Language Models (LLMs) still struggle with social reasoning, according to a new benchmarking task called the Epistemic Asymmetry Schelling Task (EAST). The EAST reveals significant gaps in LLMs' ability to navigate…
Researchers have developed a language-guided vision-AI framework for waste management that integrates visual and linguistic reasoning, achieving a BLEU score of 0.8291 and a BERTScore of 0.9273 on the newly constructed W…
Researchers have developed a multi-agent system using an expert-editor stepwise questioning strategy that significantly outperforms traditional LLM methods on long-document summarization, as demonstrated on two scientifi…
The Ministry of Education and the Korea Education and Research Information Service on Tuesday launched a student advisory group of 22 teams comprising roughly 60 immigrant-background students and 22 guiding teachers to p…
A new voice anonymization model from researchers using a frozen wav2vec2 encoder and HiFi-GAN vocoder achieves a 2.53% word error rate while preserving content and emotion, with an equal error rate of 13.39 in the Voice …
A new framework called Self-Verified Reasoner (SVR-R1) allows AI models to self-verify their answers by asking themselves 'Yes or No' before finalizing a response, boosting accuracy on vision-language reasoning tasks wit…
The Education Ministry and the Korea Education and Research Information Service launched an advisory group of 60 migrant-background students from 15 countries in June to improve Korean for Everyone, the government's AI-p…
A developer shared techniques for making AI-written content sound more human, including a burstiness pass to vary sentence length, a contraction replacement engine, injecting specific numbers, and a banned words sweep of…