Natural Language Processing news and analysis on Web Pulse: 5581 curated articles tracking the latest Natural Language Processing developments, tools, and research, updated continuously from vetted sources.
A technical walkthrough explains how GPT-2's decoder processes a 768-wide vector through 12 layers to predict the next word, detailing every matrix multiplication and parameter count without hand-waving. The article demo…
Natural Language Processing (NLP) is transforming business operations by converting unstructured text into actionable data, with the global NLP market valued at $59.7 billion in 2024. Key capabilities include sentiment a…
Meta has prototyped "super sensing" smart glasses that continuously capture audio and photos every few seconds, enabling AI-powered recall of the wearer's day, according to the Financial Times. The company is considering…
Rubén González released LangDrift, an open-source tool for testing whether AI agents preserve behavior across languages, including tool selection and arguments. The tool runs YAML scenarios against any HTTP agent endpoin…
A developer built an AI-powered reverse hiring platform using Next.js, Supabase, and OpenAI. The platform flips the traditional hiring model by requiring companies to find candidates, disclosing salary upfront, and keepi…
A developer built Ved, a fully offline, self-hosted AI voice assistant that runs on consumer hardware with 12GB VRAM. The project aims to evolve into a server-client architecture for smart home integration and remote acc…
Decispher launches a system of record for engineering teams that captures decisions, conventions, and constraints from Slack, GitHub, and Jira conversations, serving them to both humans and AI agents. The platform uses a…
A developer built a Slack-based AI assistant using Open Claw, an AI agent framework, to execute natural language commands through tool chaining. The system integrates session memory and persistent message cursors to avoi…
A Nature study published in May 2026 found that commercial large language models including ChatGPT and Claude gave more favorable answers about China when prompted in Chinese than in English, with human raters judging Ch…
A developer explains how prompt engineering can dramatically improve AI responses, with techniques like role-based prompting, chain-of-thought reasoning, and format specification. The post emphasizes that mastering promp…
South Korea's Ministry of Science and ICT is scaling an AI dubbing program after AI-localized K-FAST channels drew about 100 million cumulative viewers across 22 countries in five months. The initiative uses voice separa…
Developers building a Slack and Teams scheduling agent discovered that large language models struggle to reason about time, even when given the current time. They fixed the issue by precomputing relative time labels (pas…
A study on text-to-SQL systems found that verification-based signals, such as LLM judges, outperform black-box signals like self-consistency in predicting query correctness, with a two-provider ensemble achieving 0.82 AU…
Researchers found substantial cross-lingual disparities in healthcare LLM Q&A, with models aligning more with English Wikipedia even for non-English prompts. Providing non-English contextual excerpts improved factual ali…
Researchers introduced SynthAVE, a large-scale benchmark for synthetic labeling of e-commerce attribute extraction, covering 12,726 products across 229 product types, 792 attributes, and four languages. They validated th…
Researchers developed SAMPA, a Whisper-based segmenter that transcribes Brazilian Portuguese speech while inserting markers for terminal prosodic boundaries. Fine-tuned on the NURC-SP dataset, SAMPA achieved competitive …
Researchers propose R^3, a framework for rectifying textual violations in video advertisements while preserving original semantic intent. The system uses a group-relative compliance experience extractor and curriculum re…
Researchers evaluated the relevance of several RAG metrics by comparing scores from four libraries (Ragas, DeepEval, RAGChecker, Opik) against human evaluators and standard metrics using a question-answering dataset. The…
Researchers studying harmful online communication in Discord cybercrime communities found that interpretation difficulty stems from slang, coded terms, and community-specific expressions. Their experiments showed that lo…
Researchers developed a benchmark to evaluate vision-language models' ability to use spatial deictic expressions across four languages. Tests showed that current models use demonstratives like "this" and "that" different…