Deep Learning in the Wild
summary
Journey from biomedical handcrafted features to production exposes three concrete gaps: models hailed by benchmarks often fail on field data, an mean average precision on COCO number can be misleading, and inference time on GPU ignores deployment constraints like mobile CPUs or motors. Real projects force attention to latency, memory and integration, reveal architectures that underperform on the actual dataset, and demand a separate large fresh test batch collected from the customer.
AI-written summary · about
chapters (4)
related talks
58mDeep Learning in Real Life
33mOvercoming the Barriers to Deep Learning in Production
25mDeep Learning in Biomedicine: Challenges and Opportunities
28mChallenges & future trends of Deep Learning
41mDeep Learning to Solve Challenging Problems
33mRunning Machine Learning in Production - a Journey to Success
01h 00mDeep Learning Revolution: Rethinking Machine Learning Pipelines
41mBeyond the Algorithm: What Makes Machine Learning Work?