The Thesis Review [19] Dumitru Erhan - Understanding Deep Architectures and the Effect of Unsupervised Pretraining
Feb 19, 2021
Dumitru Erhan, a Research Scientist at Google Brain, dives into the fascinating world of neural networks. He discusses his groundbreaking PhD work on deep architectures and unsupervised pretraining. The conversation touches on the evolution of deep learning, the significance of regularization hypotheses, and the philosophical nuances in AI task conceptualization. Dumitru shares insights into the transition from traditional computer vision to deep neural networks and highlights the importance of unexpected outcomes in enhancing research understanding.
AI Snips
Chapters
Books
Transcript
Episode notes
Accidental Entry to ML
- Dumitru Erhan's entry into machine learning was accidental, starting with an internship in Helsinki, Finland.
- Despite no prior knowledge, he found it intriguing and pursued it further at his university with Herbert Jaeger.
A Risky PhD Topic
- Yoshua Bengio initially hesitated to take Dumitru as a PhD student due to deep learning's uncertainty in 2006.
- Dumitru accepted the risk, recognizing the field's exciting potential.
Exploring the Unknown
- The allure of exploring the unknown in deep learning excited Dumitru.
- The techniques worked, but the underlying reasons were unclear, creating a fertile ground for research.
