Lightning Talks
summary
Brief lightning sequence lists concrete claims and demos: a mammogram pipeline uses candidate extraction plus neural nets that now match radiologist performance on calcifications; a real‑time gesture system trains on 120,000 Mechanical Turk videos with 27 classes and 3D convolutions but the stage demo is flaky; a quick CNN on CIFAR‑10 hits ~50% in ~10 minutes; medical concept vectors come from ontology random walks; Keras tutorials and an LSTM lyric generator round out the set. Pick talks where numbers, code, and working demos matter.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 2:58PyCon DE in Karlsruhe
- 8:12Python turning into Java
- 12:14EuroSciPy in Erlangen
- 16:22Breast cancer screening ML
- 21:52Deep learning for beginners
- 22:41Medical concepts representations
- 29:46Gesture recognition demo
- 33:58Nobel Prize with AI
