Kotlin strives for Deep Learning
summary
Walks through building a from‑scratch 3‑input XOR neural net in Kotlin on the JVM, initializing weights with a fixed seed, using sigmoid activation and training for 1,000 epochs where error drops from ~0.34 to ~0.03 and the model generalizes to unseen inputs. Covers practical tooling: Komputation (Kotlin) with CUDA/NVIDIA GPU support and a Kotlin NLP library for higher‑level tasks, plus notes on matrix ops, backprop derivatives and production reuse of learned weights.
AI-written summary · about
chapters (7)
- 0:00Introduction and motivation
- 6:48Deep learning fundamentals
- 11:06Why Kotlin for AI
- 13:58Coding exercise neural network
- 34:44Production frameworks and next steps
- 40:15Takeaways and future outlook
- 41:45Q&A session
related talks
49mDeep Dive into Deep Learning with KotlinDL
58mTwo Years in the Life of a Deep Learning Library in Kotlin
51mMachine Learning Exposed Deep and Reinforcement Learning
44mMachine Learning Exposed
47mDeep Learning: a practical introduction from an Android dev perspective
29mGradient descent is taking away your Android developer job
43m(Deep) Learn You a Neural Net for Great Good!
37mCode First, Math Later: Learning Neural Nets Through Implementation and Examples