Natural Language Processing news and analysis on Web Pulse: 3971 curated articles tracking the latest Natural Language Processing developments, tools, and research, updated continuously from vetted sources.
A new benchmark and baseline study evaluating retrieval-augmented generation (RAG) for French immigration law finds that retrieval improves administrative guidance at both model scales tested. Comparing a parametric LLM …
KT and Amorepacific jointly developed Lemon (Lab Efficiency Mode ON), an AI assistant for researchers, as part of a data integration project called Data Highway that reorganized 70 years of Amorepacific's R&D data into a…
Fetch built FAST, a self-serve conversational BI agent, on ClickHouse Cloud, which ingests over 150 billion rows a day from receipt and offer data. The platform lets non-technical teams query purchase data in plain langu…
Myriad Genetics rebuilt its AI platform after costs ballooned, and now processes documents in 20 seconds instead of 10 minutes, accelerating repayment cycles. The Salt Lake City-based molecular diagnostic testing company…
A developer reports that traditional SEO tactics like keyword density and link building are losing effectiveness as AI-driven search prioritizes clear, direct answers. The developer now writes every claim as if it could …
A developer observed a customer's RAG application producing confident but wrong answers, tracing the problem to poor retrieval rather than the LLM. The developer argues that most RAG quality issues stem from retrieval, p…
A developer explores lossless compression strategies to maximize information density within LLM context windows for RAG agents, improving performance and reducing costs. Techniques include structured entity extraction an…
Lorraine, an open-source Flutter meeting recorder for macOS, captures system audio and microphone, transcribes meetings locally or via a Modal deployment running WhisperX, and recognizes speakers across sessions using vo…
A Retrieval-Augmented Generation (RAG) application integrates a vector database with a Large Language Model (LLM) to provide AI access to external datasets, operating in indexing and inference phases. The pipeline requir…
Fabrice Niyigaba ’27 led a study creating KinyaProp, the first digital tool for detecting online propaganda in Kinyarwanda, the national language of Rwanda’s 15 million people, and possibly the first for any Bantu langua…
TopTierClass built an AI system that reads and responds to Google reviews, tackling challenges like constrained generation and sentiment analysis. The system uses Gemini for LLM responses, treats brand voice as hard cons…
A content workflow expert outlines seven feedback loops for self-improving AI content pipelines, including an upstream filter loop that evaluates briefs before writing and a retrieval refinement loop that checks source q…
Google is developing a Live Translate widget for its Google Translate app, allowing users to start live translations without opening the app, based on an APK teardown of version 10.28.62.953449214.8-release. The widget c…
Tencent's workplace communication platform Enterprise WeChat has opened beta testing of an AI assistant named Dayuan, which users can activate by swiping left from any interface to access documents, spreadsheets, meeting…
Hugging Face tokenizer code in the Chapter 3 course notebook fails because the Datasets library now returns a Column object instead of a plain Python list when indexing by column name, breaking the tokenizer's expected i…
AI customer service systems, which use natural language processing, machine learning, and generative AI, can automate over 70% of customer queries and reduce inbound call handling time by 38% for mature adopters, accordi…
A new open-source toolchain applies ASD-STE100 Simplified Technical English to large language model outputs, achieving a 74% reduction in writing violations on Claude Sonnet and a 50% reduction on GPT-5.5, according to t…
A developer explains that chain-of-thought prompting works because it effectively increases the model's compute budget by turning a single forward pass into multiple passes, using written tokens as external memory. The t…
A document intelligence platform built on Snowflake using Cortex AI functions enables enterprises to parse, classify, extract, summarize, and search PDFs and images natively within Snowflake's governance perimeter, proce…
AI chatbots like ChatGPT, Gemini, and Claude often produce confidently wrong answers due to their underlying architecture of next-token prediction and reinforcement learning from human feedback (RLHF), which prioritize f…