CS359A: Research seminar on complexity theory
General Information
Seminar description
A research seminar on concrete complexity theory.
We will study powerful and versatile techniques for analyzing basic models of computation — decision trees, DNF formulas, halfspaces, and juntas — and see applications of these techniques to a variety of areas including circuit lower bounds, derandomization, quantum complexity, learning theory, and applied probability.
A major focus of the seminar will be on mapping out the research frontier and developing fruitful research directions. In the final few meetings of the quarter we will tackle research problems together.
We will study powerful and versatile techniques for analyzing basic models of computation — decision trees, DNF formulas, halfspaces, and juntas — and see applications of these techniques to a variety of areas including circuit lower bounds, derandomization, quantum complexity, learning theory, and applied probability.
A major focus of the seminar will be on mapping out the research frontier and developing fruitful research directions. In the final few meetings of the quarter we will tackle research problems together.
Topics covered
- Decision trees
Query complexity and relationships among complexity measures
Every decision tree has an influential variable
The Aaronson-Ambainis conjecture and the power of quantum computation
P = NP ∩ coNP for decision trees
The sensitivity theorem
- DNF formulas
Random restrictions and switching lemmas
Linial-Mansour-Nisan: Fourier spectrum of DNF formulas
Mansour's conjecture and its implications
Razborov-Smolensky polynomial approximators
Bazzi's theorem: Bounded independence fools DNFs
Fredman-Khachiyan and the monotone dualization problem - Halfspaces
Noise sensitivity of halfspaces
Halfspaces and the Central Limit Theorem
Mossel-O'Donnell-Oleszkiewicz Invariance Principle
The Gotsman-Linial conjecture
Bounded independence fools halfspaces - Juntas
Learning in the presence of irrelevant information
Mossel-O'Donnell-Servedio
Valiant's light bulb problem
Seminar format
• First few meetings: Li-Yang will give an overview of the topics listed above
• Bulk of the quarter: Student-led deep dives into specific topics and results
• Final few meetings: Research discussions
• Bulk of the quarter: Student-led deep dives into specific topics and results
• Final few meetings: Research discussions
Evaluation
(Tentative; more details to follow.)
• Discussion lead for 1-2 meetings
• Short summaries of meetings
• Class participation
• Short summaries of meetings
• Class participation
(Seminar meetings will not be recorded.)