The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) *Bonus Episode* A Quantum Machine Learning Algorithm Takedown with Ewin Tang - TWiML Talk #246
Apr 1, 2019
In this engaging conversation, Ewin Tang, a PhD student at the University of Washington, takes listeners on a journey through her groundbreaking work on quantum-inspired algorithms for recommendation systems. She reveals how her revolutionary approach challenges traditional notions about the necessity of quantum computing. Ewin dives into the intricacies of quantum superposition, discussing both its potential advantages and the complexities of real-world application. With skepticism swirling in the research community, she advocates for a critical examination of quantum claims.
AI Snips
Chapters
Transcript
Episode notes
Quantum Algorithm Takedown
- Ewin Tang's thesis challenges the perceived advantage of quantum machine learning algorithms for recommendation systems.
- She discovered a classical algorithm that achieves similar speedups, questioning the need for quantum computing in this context.
Lower Bound Research
- Tang initially aimed to prove the quantum algorithm's exponential speed advantage.
- However, roadblocks during her research hinted at a possible classical counterpart.
Sublinear Time Computation
- Existing classical algorithms faced limitations in outputting large vectors efficiently.
- Tang's insight involved using probability distributions to represent vectors, enabling sublinear time computation.