research
Exact Learning Broke an Old Quantum-Speedup Conjecture
The new concept classes go beyond the Grover and Bernstein-Vazirani patterns.
Summary
The new concept classes go beyond the Grover and Bernstein-Vazirani patterns.
The authors construct exact-learning concept classes whose randomized and deterministic membership-query complexities have cubic-scale separations from quantum query complexity, refuting a longstanding conjecture and matching prior upper bounds up to the stated factors. The result concerns oracle query complexity, not end-to-end training time.
Why it matters
The new concept classes go beyond the Grover and Bernstein-Vazirani patterns.
Limits and context
- The result concerns oracle query complexity, not end-to-end training time.
Key claims
The new concept classes go beyond the Grover and Bernstein-Vazirani patterns.
Qualification: The result concerns oracle query complexity, not end-to-end training time.
Evidence: source-2026-09-30-020
Sources
- arXiv preprint 2609.38073arXiv · primary research
Corrections
No corrections have been recorded for this story.