Deep Learning Needs Clojure
summary
Demo uses Clojure's persistent data structures and REPL-driven workflow to improve traceability and debugging for convolutional nets; it combines spec-driven network generation with genetic programming to evolve architectures and yields roughly 90% accuracy on a standard handwritten dataset after minutes of search. Points to JVM and JS bridges (Cortex, DeepLearning4J, TensorFlow, Guilds minh) and to Graal/Truffle bringing Python to the VM, arguing Clojure tooling makes experimentation and explainability more practical.
AI-written summary · about
chapters (7)
- 0:00Start
- 13:07EARTH NEEDS CLOJURE
- 13:43DEBUGGING
- 19:10Use Genetic Programming
- 20:06clojure spec
- 22:39The Way Forward
- 26:26Learn from the Python deep learning ecosystem
related talks
28mMachine Learning Live
25mMore Clojure Less Complication
44mMachine Learning with Artificial Neural Networks in Clojure
01h 01mDeep Learning and Java
38mMachine Learning with Clojure and Apache Spark
44mDeep Learning in Production
40mLessons Learned Creating a Successful Machine Learning Startup with Clojure
26mThe Feynman Machine: AI architecture in Clojure/Neanderthal