DEV Community

Machine Learning

A branch of artificial intelligence (AI) and computer science which focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy.

Posts

👋 Sign in for the ability to sort posts by relevant, latest, or top.
Speculative decoding shifted our output distribution and evals missed it

Speculative decoding shifted our output distribution and evals missed it

1
Comments 1
4 min read
BĂątir un SystĂšme de Maintenance PrĂ©dictive : De l’IoT Industriel Ă  l’Analyse Vectorielle đŸ­đŸ€–

BĂątir un SystĂšme de Maintenance PrĂ©dictive : De l’IoT Industriel Ă  l’Analyse Vectorielle đŸ­đŸ€–

Comments
3 min read
I Gave My AI Agent a Conscience and a Council

I Gave My AI Agent a Conscience and a Council

2
Comments
5 min read
Lessons from Building an AI Video Cleanup Tool

Lessons from Building an AI Video Cleanup Tool

Comments 2
4 min read
Shared expert pool reduces parameters while maintaining performance

Shared expert pool reduces parameters while maintaining performance

Comments
3 min read
RAG SOTA: I Tested 7 Pipelines and Built SEQUOIA (Open Source)

RAG SOTA: I Tested 7 Pipelines and Built SEQUOIA (Open Source)

Comments 2
2 min read
Violence Detection in Video Files: Training AI to detect human-centric emotional responses

Violence Detection in Video Files: Training AI to detect human-centric emotional responses

Comments
2 min read
# I Built a RAG System That Enforces Its Own Citations — And Blocks Its Own Merges

# I Built a RAG System That Enforces Its Own Citations — And Blocks Its Own Merges

Comments
5 min read
Applied Scientist Skills Companies Want in 2026: A comprehensive analysis on 3,146 active postings

Applied Scientist Skills Companies Want in 2026: A comprehensive analysis on 3,146 active postings

5
Comments
17 min read
DeepSeek or Qwen 3 Max? I Ran Both for a Month on Client Work

DeepSeek or Qwen 3 Max? I Ran Both for a Month on Client Work

3
Comments 1
8 min read
Why we run two scoring tracks (LLM + Mediapipe) for our AI face-rating tool

Why we run two scoring tracks (LLM + Mediapipe) for our AI face-rating tool

Comments
3 min read
Your bandit's exploration floor probably violates its own floor

Your bandit's exploration floor probably violates its own floor

Comments
3 min read
A model with R-squared near 0 can still give valid 90% prediction intervals - here's why (and the catch)

A model with R-squared near 0 can still give valid 90% prediction intervals - here's why (and the catch)

1
Comments
4 min read
From Breaking AI Filters to Dressing Real People: A Cross-Domain Creator Worth Watching

From Breaking AI Filters to Dressing Real People: A Cross-Domain Creator Worth Watching

10
Comments 2
4 min read
Conformal prediction silently breaks under drift - and how to make it hold

Conformal prediction silently breaks under drift - and how to make it hold

Comments 2
4 min read
👋 Sign in for the ability to sort posts by relevant, latest, or top.