Super Data Science: ML & AI Podcast with Jon Krohn 721: Quantum Machine Learning, with Dr. Amira Abbas
9 snips
Oct 10, 2023 Explore Quantum Machine Learning with Dr. Amira Abbas from the University of Amsterdam. Learn about qubits, quantum entanglement, and quantum neural networks. Discover the best problems for quantum ML and recommended ML tools for quantum computing. Gain insights into the potential of quantum computing in enhancing machine learning tasks.
AI Snips
Chapters
Transcript
Episode notes
Seeing Quantum Computers In The Lab Felt Like Art And Engineering
- Amira describes seeing an IBM superconducting quantum computer as a gold chandelier requiring extreme cooling and a photonic system in a Toronto penthouse.
- The contrast highlights different hardware approaches and the expensive precision engineering behind current devices.
Quantum ML Split Into Encode Evolve Measure Steps
- Quantum machine learning breaks into three steps: encode classical data into quantum states, evolve the states (parameterized circuits), and measure to get labels.
- Amira emphasizes each step has many design choices and depends on data structure and task.
Quantum Feature Maps Offer Hard Kernels But Are Artificial
- Encoding classical data into quantum states can act as a feature map for kernel methods, giving potentially classically intractable kernels.
- Amira notes demonstrated quantum kernels often use artificial datasets; finding natural useful feature maps remains open.