Blog · Cognition · Research · Design · Artificial intelligence
Notes on people,
models and design.
Articles grounded in cognitive science, user experience research and applied artificial intelligence, explained in accessible language.
01 — BLOG
Rigorous concepts, explained without distortion.
Select one or more topics to find their intersection.
4 articles / 20
One vector space for everything? Benchmarking CLIP, FashionCLIP and OpenAI
Two catalogues, three encoders and one protocol for choosing a vector space according to the actual task.
- Search & recommendation
- Computer vision
- AI & machine learning
From embedding space to a visual-search service
The journey from CLIP experiments to a tested, observable and reproducible vector-search API.
- Search & recommendation
- Computer vision
- AI & machine learning
Designing a configurable multimodal product recommender
Architecture, incremental processing, multimodal fusion and HNSW search for recommendations without user history.
- Search & recommendation
- AI & machine learning
- Computer vision
Image segmentation: isolating the product before computing similarity
Why I explored YOLOv8 product segmentation before embedding images in a visual-search engine.
- Computer vision
- Search & recommendation
- AI & machine learning
Each post separates findings, methodological limitations and possible uses. The primary sources are my publications, doctoral research and cited papers.
Suggest a topic
A good question may deserve an article.
Psychometrics, behavioral signals, attention, user experience or applied artificial intelligence: share the subject you would like to see explored.
Suggest a topic ↗