History
#
- Idea: Learning the feature descriptors from data, rather than hand-engineering them
What it possible
#
- We have a lot of data
- The hardware is better (GPUs are used)
- Cloud computing
- Deeper -> models are more complex
Deep Learning Today
#
- Image classification
- Object detection
- -> Important because important for some idustries
- -> like Self-driving cars
- Games (AlphaGo, AlphaStar)
- Emoticon Suggestions
- Language translation
- Virtual assistants
- Medical: cancer detection, etc
Deep Learning Market & Job Perspective
#
- DL will help people do their jobs
- Whole society will benefit
- DL experts are and will be needed
- High level knowledge good, but low-level needed to stand out