RAG - AI Feta, the news about scientific AI research

tekoäly

Haku muuttuu vastaukseksi – verkko alkaa sekä etsiä että selittää yhtä aikaa

Kuvittele kirjoittavasi hakukenttään: ”Mitä minun kannattaa katsoa Tokiossa kolmen päivän matkalla?” Ennen sait listan linkkejä. Uutena vastauksena voi tulla sujuva selostus, jossa ehdotetaan reitti, avataan aukioloajat ja kerrotaan, miksi jokin kaupunginosa on iltaisin parhaimmillaan – ja teksti on koottu useista lähteistä hetki sitten. Sama palvelu saattaa vielä lopuksi tiivistää linkkien sisällön

tekoäly

Kielimallin huomio toimii yllättävän hyvin pitkien tekstien hakijana

Moni tuntee tilanteen: edessä on 180-sivuinen raportti, ja pitäisi löytää vastaus yhteen täsmäkysymykseen. Hakutoiminto löytää kymmeniä osumia, mutta oikea kohta on aina sen taulukon alaviitteessä tai liitteessä, johon teksti viittaa. Sama ongelma vaivaa myös älykkäitä keskustelubotteja. Ne lupaavat lukea pitkät tiedostot, mutta harhailevat helposti väärään kappaleeseen tai vastaavat luottavaisesti hutiin.

AI

LADFA: AI that maps your personal data in privacy policies

AI that reads privacy policies so you don't have to Privacy policies are long, legal, and hard to compare. LADFA is a new AI framework that turns those walls of text into a clear map of how companies collect, use, store, and share your personal data. How it

LLM

Private LLM Inference on Consumer Blackwell GPUs: A Practical Guide for Cost-Effective Local Deployment in SMEs

Want private AI without cloud risk or spend? This study shows SMEs can run production LLMs on NVIDIA Blackwell consumer GPUs (RTX 5060 Ti, 5070 Ti, 5090). * Cost: $0.001–$0.04 per million tokens (electricity only) — 40–200x cheaper than budget cloud APIs. * ROI: Hardware can pay for itself

Water

Meet WaterCopilot: AI for Smarter Water Decisions in the Limpopo Basin

Meet WaterCopilot, an AI assistant co-created by IWMI and Microsoft Research to help manage the Limpopo River Basin. It turns scattered documents and live hydrology feeds into clear, actionable answers. * Unifies policy files, reports, and sensor data in one chat. * Answers with transparent source links you can verify. * Provides live

AI

Tiny AIs vs. Server Logs: Who's Ready for Real Time?

System logs are messy and massive. This study tests whether small language models (SLMs) can reliably read them fast. Instead of treating "severity" labels as the end goal, the authors use severity classification as a probe of real runtime log understanding. * Best performer: Qwen3-4B hit 95.64% accuracy

AI

ArcAligner: Helping AI use compressed context without losing accuracy

LLMs with Retrieval-Augmented Generation (RAG) work better when they read lots of context — but long prompts are slow and pricey. Compressing the context helps, yet models often lose the thread and answer worse. ArcAligner is a lightweight module that helps models make sense of highly compressed context. It "aligns&

AI

From Pixels to Posts: Retrieval-Augmented Fashion Captioning and Hashtag Generation

Meet a new way to write captions and hashtags for fashion photos—grounded in what is actually in the picture. * Detects multiple garments in an image with a YOLO-based model. * Extracts dominant colors and infers fabric and gender by retrieving similar products with CLIP and FAISS. * Packages these facts as

AI

PersonaAgent with GraphRAG: AI that knows your taste

AI that understands your preferences PersonaAgent with GraphRAG is a new way to build language model agents that adapt to you. The agent learns a compact "persona" from your past choices and behaviors, stored in a knowledge graph. It also finds communities of related information across documents to

RAG

Which LLM wins at RAG Q&A?

Which LLM wins at RAG Q&A? RAG (Retrieval-Augmented Generation) helps reduce hallucinations by grounding answers in source documents. This study tested five 7B-class models on computer science literature Q&A to compare accuracy and speed. * GPT‑3.5 + RAG answered both yes/no and long-form questions effectively.