Applying Machine Learning to IoT Use Cases with Deeplearning4j
summary
Explains that Deeplearning4j targets the JVM and production use: model development with Java-based tensors, deployment on Raspberry Pi and Jetson, and Spark integration for streaming from MQTT/Kafka. Describes a working reinforcement learning traffic demo that reduces wait times, a deployed electricity-theft detector in China, and experiments with bundled CUDA from Maven so GPU setup isn’t required. Argues most practical projects are simple yes/no classifiers and stresses model versioning, retraining and data batching.
AI-written summary · about
chapters (14)
- 0:00Intro
- 6:46Context
- 11:46Transfer Learning
- 19:03Where does ML meet IoT
- 20:00ML ingest
- 20:55MQTT
- 24:19Architectures
- 30:46Traffic Flow
- 31:46Deeplearning4j
- 37:43Use Cases
- 39:16Federated Learning
- 40:21federated learning example
- 40:59Azul CTO
- 41:26Questions
related talks
35mJVM based DeepLearing on IoT data with Apache Spark
44mDeep Learning in Production
25mDeeplearning4j: How to run AI workloads on Jakarta EE compliant servers
44mDeep Learning with DeepLearning4J and Spring Boot
01h 01mDeep Learning and Java
51mGetting Started with Deep Learning for Software Developers
50mML in Java, YES it's possible!
29mAchieve real-time inference at scale on AWS with Deep Java Library