mcubed london - data science at the edge
Download as PPTX, PDF4 likes823 views
The document discusses the integration of data science technologies like NiFi and TensorFlow with a focus on processing and analyzing data at the edge. It highlights the evolution from traditional computing methods to cloud-based solutions for tasks like face detection and recognition, emphasizing the importance of context in data analysis. Various models and their operational challenges, including computational demands and the use of existing clusters, are addressed throughout.
mcubed london - data science at the edge
- 1. Data science at the Edge With NiFi, TensorFlow and a proper cluster for good measure Simon Elliston Ball @sireb
- 2. Simon Elliston Ball • Product Manager • Data Scientist • Elephant herder • @sireb
- 3. Data gravity 588,000,000 km • Size • Distance
- 4. Other types of data gravity • Compliance • Legislation • Political • Paranoia Photo: https://flic.kr/p/JvW7qh
- 5. Sampling vs Big Data: a quick history • Before we had cloud, clusters and GPUs… • MPP • Super Computers • Grids • Cut down data size to fit in memory
- 6. A quick intro to NiFi • Guaranteed Delivery • Prioritized queuing and buffering • Data provenance • Bi-directional communication • Security – Authentication and multi- role authorization • Visual command and control • Templating • Robust API
- 7. and lots of adapters
- 8. Demo: sending stuff around • Pushing camera frames to the cloud
- 9. Face detection Key point locations Lightweight models Low contextual data
- 10. face detection • Simple haarcascader in opencv: https://github.com/simonellistonball/nifi- OpenCV
- 11. Dlib Face Detection • 68 Facial Point Model • c. 100MB
- 12. Tensorflow in NiFi • Our haarcascade was… Face detection didn’t do a great job • Neural Networks • Relatively Large models • Haarcascader: 677KB of XML • Facenet trained model on LFW: 168 MB (and that’s zipped protobufs) • Tensorflow: https://github.com/tspannhw/nifi-tensorflow-processor.git
- 13. Face recognition • Huge databases of face hashes and feature measures • Extra information and context around the person • Computationally expensive and heavy network use • Apple Face ID demo… too many people had tried the device beforehand, blew the database. One or two faces is easy, millions is another matter
- 14. Rocket ship to the cloud https://www.nasa.gov/sites/default/files/thumbnails/image/s83-35620-3k.jpg
- 15. Cloud: ML all packaged up… for a price
- 16. Tensorflow on Spark • Why? • Doesn’t TensorFlow already have a distributed compute model? Existing clusters, multi-purpose clusters: • Tensorframes, TensorflowOnSpark, CaffeOnSpark, Spark ML, SQL • When? • Training, batch scoring
- 17. Broadening the example • Where is your context? • Why do you need context? • Detection • Explanation
- 18. Body worn video • Record everything • Record when you remember to press the button • Record when it matters What about? • Live assist • Evidence and accountability
- 19. Netflow Cybersecurity: progressive context • Record everything: PCAP • Send up the (maybe) interesting bits • Fetch detail on demand PCAP at Edge 1ST Pass Model Security Data Analytics Platform adds context, more compute intensive modelling etc Hmmm… That’s interesting Let me tell you more… “small” data flow
- 20. ANPR: or why you can’t hide from parking fines
- 21. Summary: progressive enhancement of context Is it worth processing? Rough-cut and hashing Expensive deep analysis @sireb 677KB of local model O(100MB) models Cloud scale models and data name Simon Elliston Ball cognitive.face.emotion surprise cognitive.face.exposure overExposure cognitive.face.noise high
Editor's Notes
- #13: And what do we get for our 180x increase in model size?