After a crazy night in the “Lab” experimenting with liquid nitrogen and all sorts of alcoholic flavours (more on that later), this is my first entry on a weekly summary of the most shocking advances in AI I had the chance to look at (not sure if I should change to daily, as almost every day there is a new publication that blows my mind away). Below my favorites for the week
The upcoming lip-reading module for “Skynet” - LIPNET: Sentence-level lipreading - (PDF)
Highlit from the abstract: LipNet is the first lipreading model to operate at sentence-level, using a single end-to-end speaker-independent deep model to simultaneously learn spatiotemporal visual features and a sequence model. On the GRID corpus, LipNet achieves 93.4% accuracy, outperforming experienced human lipreaders and the previous 79.6% state-of-the-art accuracy.
My take:
- Truly impressive progress in using Deep Learning for lip reading.
- LipNet is trained end-to-end to make speaker-independent sentence-level predictions. Look out for this to be applied in industrial spionage (Remember to cover your mouth from now on when cameras around)
Discovering new drugs with Deep Learning
Ingesting a heap of drug data allows a machine-learning system to suggest alternatives humans hadn’t tried yet. (Link, Link to Paper )
Highlits: A new artificial intelligence aims to create new drugs by learning what molecules already work and what may fit with existing patterns. “The system can dream up structures more independently of humans and without lengthy simulations. It leverages its own experience, built up by training machine-learning algorithms with data on hundreds of thousands of drug-like molecules. It explores more intuitively, using chemical knowledge it learned, like a chemist would”
My take:
- Another breakthrough in another field boosted by Deep Learning which will completely change the way we discover drugs and will accelerate the progress in this field by reducing time and cost and will lead to revolutionary discoveries in key areas like cancer research etc.
- This alongside other recent breakthroughs like new gene editing techniques like CRISPR-Cas9 will transform the face of medicine forever - Think like Medicine 2.0
There are a couple of more interesting developments in AI this week but I will keep them for my next post…