sateesh
Joined 17 years ago
Karma
521
Submissions
268
Member since
Aug 10
sateeshkumarb AT yahoo.com
View on Hacker NewsJoined 17 years ago
Multimodal retrieval does not suit this domain. CLIP-style embeddings wash out exactly the fine detail that matters in charts, tables, and annotated screenshots, and short technical queries ("how do I configure X") give too little signal to match against image vectorsTo validate the approach I trained the model on generating synthetic data. I did look at datasets available at: https://github.com/logpai/loghub, https://www.unb.ca/cic/datasets/index.html but couldn't find one that would suit my needs.
The approach seems to work on synthetic dataset (with ROC AUC score: 0.9957) but couldn't try it out in a real world dataset. Seeking feedback on the approach.
Understanding Raft can be tough. In fact, I’ve seen conversations recently on social media in which actual technical leaders of infrastructure companies demonstrate a lack of understanding (!). Point being, you’re not alone.
Anyway I now notice that the article was written in 2023, probably I'm being too pedantic.