~/conference

ITCS 2021

theoretical computer scientists, complexity theorists, quantum computing researchers, algorithm designers, machine learning researchers, distributed systems researchers, market design researchers

planted clique hypothesis; trace reconstruction algorithms; list-recovery threshold computation; QBF CDCL versus Q-resolution; linear cross-entropy benchmarking limits; hardness of training ReLU networks; pseudorandom generators for permutation branching programs; distributed load balancing frameworks; lower bounds for off-chain plasma protocols; interactive proofs for verifying machine learning

titleauthorlength
The entropy of lies: playing twenty questions with a liarYuval Dagan, Yuval Filmus, Daniel Kane, Shay Moran23m
Distributional 20 Questions with up to K lies requires on average H(mu) + K·H2(mu) yes/no queries, where H(mu) + K·H2(mu) is tight (matching lower bound up to…
Talk
Comparing computational entropies below majority (or: When is the dense model theorem false?)Russell Impagliazzo, Sam McGuire39m
Shows explicit separations proving the dense model theorem false below majority: constructions of delta-dense distributions that are pseudorandom for small…
Talk
Algorithmic Persuasion with EvidenceMartin Hoefer, Pasin Manurangsi, Alexandros Psomas28m
Paper at ITCS 2021 studies persuasion with unforgeable evidence, gives a polynomial-time max-flow algorithm that computes a sequential equilibrium, and proves…
Talk
The Complexity of Finding Fair Independent Sets in CyclesIshay Haviv19m
Shows that finding an independent set in a cycle that picks about half from each given part is PPA-complete, and that the version asking for two disjoint such…
Talk
Sharp Threshold Rates for Random CodesVenkatesan Guruswami, Jonathan Mosheiff, Nicolas Resch, Shashwat Silas, Mary Wootters16m
Characterizes exact threshold rates for a wide class of monotone symmetric properties of completely random codes: R = 1 − max_tau H_q(tau)/b, shows these…
Talk
Pseudorandom Generators for Unbounded-Width Permutation Branching ProgramsWilliam Hoza, Edward Pyne, Salil Vadhan26m
Presents an explicit INW-style generator that has seed length O(log n log(1/ε) + log n log log n) for permutation branching programs of arbitrary width, proved…
Talk on
Pipeline InterventionsEshwar Ram Arunachaleswaran, Sampath Kannan, Aaron Roth, Juba Ziani25m
Layered graphical model where agents follow stochastic transition matrices and a central planner pays convex costs to change those matrices under budget B…
Talk on
A Polynomial Degree Bound on Equations for Non-rigid Matrices and Small Linear CircuitsMrinal Kumar, Ben Lee Volk27m
Paper proves an explicit existential bound: for non-rigid n-by-n matrices with parameters εn and εn^2 there exists an equation of degree O(n^3) vanishing on…
Talk
The Strongish Planted Clique Hypothesis and Its ConsequencesPasin Manurangsi, Aviad Rubinstein, Tselil Schramm24m
Postulates that for kappa = N^{0.1} planted clique cannot be solved in time N^{o(log N)} and uses a randomized graph product that samples N random L‑vertex…
Talk
Sample efficient identity testing and independence testing of quantum statesNengkun Yu09m
Constructs an explicit independent measurement M on a D-dimensional state whose post-measurement classical distributions satisfy ||P-Q||_2 =…
Talk on
Understanding the Relative Strength of QBF CDCL Solvers and QBF ResolutionOlaf Beyersdorff, Benjamin Böhm28m
Paper proves that standard QBF CDCL and Q-resolution are incomparable by giving two explicit families: QParity has polynomial-size QCDCL refutations but…
Talk
The Quantum Supremacy Tsirelson InequalityWilliam Kretschmer20m
Proves a query lower bound for linear cross-entropy benchmarking: any polynomial-time quantum algorithm needs about 2^{n/4} queries to beat the trivial quantum…
Talk on
A New Model for Ant Trail Formation and its Convergence PropertiesMoses Charikar, Shivam Garg, Deborah M. Gordon, Kirankumar Shiragur33m
A flow-based pheromone model on directed graphs is defined with decay, node leakage and forward/backward unit flows and a linear decision rule; for two…
Talk on
Approximately Strategyproof Tournament Rules in the Probabilistic SettingKimberly Ding, S. Matthew Weinberg23m
Shows that for independent, weakly ε-bounded probabilistic tournaments two natural recursive rules—randomized deathmatch and randomized single‑elimination…
Talk
Even the Easiest(?) Graph Coloring Problem is not Easy in Streaming!Anup Bhattacharya, Arijit Bishnu, Gopinath Mishra, Anannya Upasana36m
Establishes one-pass streaming bounds for estimating the number of monochromatic edges given a coloring: an algorithm that stores sample R of size n/√T (first…
Talk
The Variable-Processor Cup GameAlek Westover, William Kuszmaul19m
Shows that allowing P to vary changes guarantees drastically: the greedy emptier remains optimal versus deterministic adversaries but its backlog jumps from…
Talk
Complexity measures on symmetric group and beyondNeta Dafni, Yuval Filmus, Noam Lifshitz, Nathan Lindzey, Marc Vinyals33m
Defines an abstract simplicial complex framework that extends degree, decision tree complexity and sensitivity from the hypercube to domains like the symmetric…
Talk
Comparison Graphs: a Unified Method for Uniformity TestingUri Meir36m
Defines comparison graph as which sample pairs to compare and proves a structural theorem: if the graph has enough edges and few two-paths dependencies then…
Talk
Circular Trace ReconstructionShyam Narayanan, Michael Ren30m
Problem: reconstruct a circular binary string from independent deletion-channel traces (deletions prob q). Main claims: worst-case reconstruction uses…
Talk
Self-testing of a single quantum device under computational assumptionsTony Metger, Thomas Vidick23m
Constructs a protocol that self-tests an EPR pair and the device’s measurements for a single monolithic quantum device under the computational assumption that…
Talk on
Polynomial-time trace reconstruction in the low deletion rate regimeXi Chen, Anindya De, Chin Ho Lee, Rocco A. Servedio, Sandip Sinha23m
Claim: a polynomial-time algorithm reconstructs any n-bit string with high probability using polynomially many traces when the deletion rate δ < n^{-1/3+ε}…
Talk
Metrical Service Systems with TransformationsSebastien Bubeck, Niv Buchbinder, Christian Coester, Mark Sellke24m
Defines Metrical Server Systems with Transformations: each request is a map from domain AT to image BT and the algorithm pays only to move into AT. Shows any…
Talk
Tight Hardness Results for Training Depth-2 ReLU NetworksDaniel Reichman, Pasin Manurangsi, Subhi Goel, Adam R. Klivans32m
Proves concrete limits and matching algorithms for depth-2 ReLU training: shows single-ReLU NP-hard (agnostic) and NP-hardness for any K≥2 in the realizable…
Talk on
A Largish Sum-of-Squares Implies Circuit Hardness and DerandomizationPranjal Dutta, Nitin Saxena, Thomas Thierauf28m
If an explicit univariate family has support sum ≥ D^{1/2+ε} then VP ≠ VNP; a constant ε yields an exponential separation. The proof uses depth reduction…
Talk
Circuit Depth ReductionsAlexander Golovnev, Alexander S. Kulikov, Ryan Williams34m
Proves structural depth reductions: any size-S circuit is an OR of 2^{S/3} formulas (toy lemma) and the main result gives a depth-3 with top fan-in 2^{S/3.9}…
Talk
Dynamic inference in probabilistic graphical modelsWeiming Feng, Kun He, Xiaoming Sun, Yitong Yin23m
Presents an algorithm that updates samples and estimates after small edits to a graphical model with cost roughly proportional to the edit size rather than…
Talk on
Theorems of KKL, Friedgut, and Talagrand via Random Restrictions and Log-Sobolev InequalityEsty Kelman, Subhash Khot, Guy Kindler, Dor Minzer, Muli Safra35m
Proves KKL, Friedgut and Talagrand without hypercontractivity by combining random restrictions with a version of the log-Sobolev inequality. Shows that…
Talk
On Rich 2-to-1 GamesMark Braverman, Subhash Khot, Dor Minzer23m
Defines an r-wise rich two-to-one label-cover variant where each left vertex induces a near-uniform distribution over 2-sized partitions of the left alphabet…
Talk
Distributed Quantum Proofs for Replicated DataPierre Fraigniaud, François Le Gall, Harumichi Nishimura, ami Paz20m
Claim: replicated-data consistency can be certified with log N qubits per certificate and per message, versus Θ(N) classically. Construction uses quantum…
Talk on
On Basing Auxiliary-Input Cryptography on NP-hardness via Nonadaptive Black-Box ReductionsMikito Nanashima29m
Shows that non-adaptive black-box reductions from NP-hard languages to an auxiliary-input pseudorandom generator would collapse the polynomial hierarchy, using…
Talk on
Spoofing Linear Cross-Entropy Benchmarking in Shallow Quantum CircuitsBoaz Barak, Chi-Ning Chou, Xun Gao33m
Claims a classical algorithm that uses light-cone marginals to sample from shallow random circuits, computes each marginal in poly(n)·2^L time where L is the…
Talk on
On the complexity of isomorphism problems for tensors, groups, and polynomials I: Tensor IsomorphismJoshua A. Grochow, Youming Qiao31m
Proves that for any constant D≥3, D-tensor isomorphism reduces in polynomial time to 3-tensor isomorphism, so a large cluster of problems (cubic forms, algebra…
Talk
Bounds on the QAC0 Complexity of Approximating ParityGregory Rosenthal17m
Constant-depth QAC0 circuits can approximate parity arbitrarily well but at steep cost: depth seven requires exponential size, and the paper gives a depth-11…
Talk on
Query complexity lower bounds for local list-decoding and hard-core predicates (even for small rateNoga Ron-Zewi, Ronen Shaltiel, Nithin Varma21m
Proves query lower bounds for local list decoding of very low‑rate binary codes and shows these bounds hold even for exponentially large lists. Main…
Talk
Is the space complexity of planted clique recovery the same as that of detection?Jay Mardia25m
Asks whether planted clique recovery can match detection in space and gives a near-complete constructive answer: a derandomized, combinatorial algorithm is…
Talk
Algorithms and Hardness for Multidimensional Range Updates and QueriesJoshua Lau, Angus Ritossa39m
Shows that multidimensional range updates flatten via a kd-style reduction to 1D giving ~Õ(√p) per operation and proves matching conditional lower bounds (no…
Talk
Two combinatorial MA-complete problemsDorit Aharonov, Alex B. Grilo28m
Two new combinatorial problems are shown MA-complete: SetCSP is MA-hard for inverse-polynomial ε, and an implicit-graph problem ACAC is in MA via random-walk…
Talk
Delegated Stochastic ProbingCurtis Bechtel, Shaddin Dughmi25m
Defines a game-theoretic model of delegated stochastic probing and gives an explicit single-proposal mechanism that is as powerful as general signaling…
Talk
Explicit SoS lower bounds from high-dimensional expandersIrit Dveer Dinur, Yuval Filmus, Prahladh Harsha, Madhur Tulsiani34m
Constructs deterministic, polynomial-time Max-3-XOR instances from Ramanujan complexes whose true optimum is at most 1 − μ (μ a fixed constant, e.g. ~0.1) yet…
Talk
Sampling Arborescences in ParallelNima Anari, Nathan Hu, Amin Saberi, Aaron Schild26m
Claims an RNC sampler that outputs a rooted arborescence from a weighted digraph with probability proportional to product of edge weights, running in polylog…
Talk
Lower Bounds on the Running Time of Two-Way Quantum Finite Automata and Sublogarithmic-Space QuantumZachary Remscrim29m
Proves concrete limits: gives a general lower bound on 2QCFA running time in terms of Dwork–Stockmeyer non-regularity D_L, shows palindrome needs exponential…
Talk on
On the Complexity of #CSP^dJiabao Lin20m
Proves a full dichotomy for Counting-CSP^d: for any finite function set F over a finite domain the partition function is computable in polynomial time exactly…
Talk
Interactive Proofs for Verifying Machine LearningJonathan Shafer, Guy N. Rothblum, Shafi Goldwasser, Amir Yehudayoff46m
Defines PAC verification (verifier with sample access plus an interactive prover) and gives concrete separations: realizable thresholds admit order 1/ε…
Talk on
Shrinkage under random projections, and cubic formula lower bounds in AC^0Yuval Filmus, Or Meir, Avishay Tal23m
Proves an n^3 lower bound on De Morgan formula size for an explicit AC^0 function built by composing a small hard gadget F0 with a surjectivity block, using…
Talk
Ordered Graph Limits and Their ApplicationsOmri Ben-Eliezer, Eldar Fischer, Amit Levi, Yuichi Yoshida28m
The paper defines a new limit object called an orderon (a function on R^2→[0,1] mixing an order coordinate and a variability coordinate), introduces a…
Talk
Error Correcting Codes for Uncompressed MessagesOfer Grossman, Justin Holmgren19m
Defines codes for polynomial-time recognizable message sets S and proves for up to 3% adversarial errors an efficiently encodable/decodable randomized code…
Talk on
Total Functions in the Polynomial HierarchyDaniel Mitropolsky, Christos Papadimitriou, Robert Kleinberg35m
Defines a hierarchy TFΣi of total search classes and proves specific natural problems sit at low levels: the empty problem is in TFΣ2 via a union‑bound /…
Talk
Relaxing Common Belief for Social NetworksNoah Burrell, Grant Schoenebeck24m
Paper defines common P_mu belief (a factional-belief relaxations of common belief) and proves it undergirds partial approximate coordination in threshold…
Talk
Tiered Random Matching Markets: Rank is Proportional to PopularityItai Ashlagi, Mark Braverman, Clayton Thomas, Geng Zhao21m
Man-proposing deferred acceptance runs in Θ(n log n) proposals and yields average ranks that scale inversely with public scores: expected rank for an agent is…
Talk
Black-Box Uselessness: Composing Separations in CryptographyGeoffroy Couteau, Pooya Farshim, Mohammad Mahmoody37m
Defines a composable notion called black-box uselessness and proves that infinitely-often one-way functions are useless for building key agreement when one…
Talk on
Pseudobinomiality of the Sticky Random WalkVenkatesan Guruswami, Vinayak Kumar28m
Defines a sticky random walk (first bit uniform, later bits stay with prob (1+λ)/2) and proves the Hamming-weight distribution differs from Bin(n,1/2) by total…
Talk
Robust Quantum Entanglement at (nearly) Room TemperatureLior Eldar16m
Constructs a family of local Hamiltonians whose thermal state remains hard to prepare: the Hamiltonian is topologically ordered and its thermal state has…
Talk on
Time-Space Lower Bounds for Proof Systems with Quantum and Randomized VerifiersAbhijit Mudigonda, Ryan Williams34m
Proves a concrete lower bound: there is a language in QCMA running in linear time on a QRAM that cannot be decided by small‑space random‑access machines in…
Talk on
Online Search With a HintSpyros Angelopoulos22m
Explores linear search with untrusted hints and gives concrete tradeoffs: exact-position hints admit a shrunk exponential strategy that is Pareto-optimal…
Talk
Non-quasi-linear Agents in Quasi-linear MechanismsMoshe Babaioff, Richard Cole, Jason Hartline, Nicole Immorlica, Brendan Lucier30m
Mechanisms that are truthful for quasi-linear agents remain structured when agents have convex payment disutility: each best response is a uniform scaling best…
Talk
Sequential Defaulting in Financial NetworksPál András Papp, Roger Wattenhofer25m
Paper models discrete-time updates of announced recovery rates in networks with debts and CDSs and shows concrete, surprising behaviors: small examples produce…
Talk
No quantum speedup over gradient descent for non-smooth convex optimizationAnkit Garg, Robin Kothari, Praneeth Netrapalli, Suhail Sherif24m
Proves that quantum algorithms give no asymptotic advantage for non‑smooth 1‑Lipschitz convex optimization in high dimension by showing a matching 1/epsilon^2…
Talk on
Quantum versus Randomized Communication Complexity, with Efficient PlayersUma Girish, Ran Raz, Avishay Tal21m
Shows a concrete partial function (a XOR lift of Aaronson’s Fourrelation) that has an efficient small-cost quantum simultaneous protocol with a few EPR pairs…
Talk on
Agnostic learning with unknown utilitiesKush Bhatia, Peter L. Bartlett, Anca Dragan, Jacob Steinhardt18m
Framework studies agnostic supervised learning where the utility U is arbitrary and only the hypothesis class F is restricted, and it introduces a K-comparison…
Talk on
On Distributed Differential Privacy and Counting Distinct ElementsLijie Chen, Badih Ghazi, Ravi Kumar, Pasin Manurangsi27m
Shows that distinct counting under constant-epsilon, delta = 1/poly(n) privacy is basically impossible in the local model: any local protocol needs error ≈ N /…
Talk on
A Generalized Matching Reconfiguration ProblemNoam Solomon, Shay Solomon33m
Shows a linear-time procedure that reconfigures any unweighted source matching into any target matching while keeping intermediate matchings within one of the…
Talk
Sensitivity Analysis of the Maximum Matching ProblemYuichi Yoshida, Samson Zhou22m
Gives a randomized algorithm that achieves a 1+ε approximation with O(1) worst-case edge sensitivity (uses randomized greedy + McGregor’s layered…
Talk
Computational Complexity of the Hylland-Zeckhauser Scheme for One-Sided Matching MarketsVijay V. Vazirani, Mihalis Yannakakis32m
Shows that computing exact HZ equilibria requires nonlinear fixed points by giving a unique irrational 4×4 instance (root of 17) and proves exact computation…
Talk
An O(n) time algorithm for finding Hamilton cycles with high probabilityRajko Nenadov, Angelika Steger, Pascal Su24m
Claim: an algorithm finds a Hamilton cycle in G(n,p) with high probability in O(n) time under the model that neighbor queries return a uniform random neighbor…
Talk
Communication memento: Memoryless communication complexitySrinivasan Arunachalam, Supartha Podder33m
Defines memoryless communication complexity where messages are a rewritable register and cost is the largest message size, proves tight equivalences with…
Talk
Relative Lipschitzness in Extragradient Methods and a Direct Recipe for AccelerationMichael B. Cohen, Aaron Sidford, Kevin Tian30m
Introduces a concrete condition called relative Lipschitzness between operator G and regularizer R and proves that standard extragradient (Mirror Prox)…
Talk
Training (Overparametrized) Neural Networks in Near-Linear TimeBinghui Peng, Zhao Song, Jan van den Brand, Omri Weinstein19m
Proves that training two‑layer ReLU networks with width m and n samples runs in near‑linear time by replacing the costly J J^T computation (previously Ω(m…
Talk on
A New Connection Between Node and Edge Depth Robust GraphsJeremiah Blocki, Mike Cinkoske23m
Gives a concrete reduction that turns an E-edge-depth-robust DAG with m edges into an E/2 node-depth-robust DAG with O(m) nodes and constant indegree by…
Talk on
Towards local testability for quantum codingAnthony Leverrier, Vivien Londe, Gilles Zémor32m
Construction produces an LDPC quantum code (hemi-cube) that encodes one logical qubit with minimum distance distance ~√n and local-testability/soundness…
Talk on
Complete Problems for Multi-Pseudodeterministic ComputationsPeter Dixon, A Pavan, N. V. Vinodchandran26m
Establishes that collision probability estimation is complete for turning approximation and multi-valued search algorithms into pseudodeterministic ones: if…
Talk on
Online Paging with a Vanishing RegretYuval Emek, Shay Kutten, Yangguang Shi30m
Online paging with cache size K and universe N over T rounds becomes achievable with predictions: a simulation algorithm SIM algorithm replaces true…
Talk on
Unknown I.I.D. Prophets: Better Bounds, Streaming Algorithms, and a New ImpossibilityJosé Correa, Paul Dütting, Felix Fischer, Kevin Schewior, Bruno Ziliotto34m
Revisits the unknown-IID prophet problem and gives three concrete results: a class of MRS algorithms analyzed via variational calculus that raises the large-n…
Talk
Differentially Oblivious Turing MachinesIlan Komargodski, Elaine Shi27m
Claim: any multi-tape Turing machine admits a compiler to a Differentially Oblivious Turing Machine that hides any single differing step while incurring only…
Talk on
Quantitative Correlation Inequalities via Semigroup InterpolationShivam Nadimpalli, Rocco A. Servedio, Anindya De32m
Proves a quantitative Gaussian correlation bound: for symmetric convex sets K,L the correlation ≥ φ(Σ_i Inf_i(K)·Inf_i(L)) with Inf_i defined as the negated…
Talk on
Shrinkage of Decision Lists and DNF FormulasBenjamin Rossman22m
Result gives a concrete shrinkage bound: the expected decision list size under a p‑random restriction is at most M^{α(p)} where α(p) is an explicit function of…
Talk
Block Rigidity: Strong Multiplayer Parallel Repetition implies Super-Linear Lower Bounds for TuringKunal Mittal, Ran Raz24m
Claims that a strong multiplayer parallel-repetition theorem (conjectured decay Val(G^n) ≤ 2^{-Ω(KN)} for the independent K-player function game) yields…
Talk on
Lower Bounds for Off-Chain Protocols: Exploring the Limits of PlasmaStefan Dziembowski, Grzegorz Fabiański, Sebastian Faust, Siavash Riahi26m
Paper proves there is no perfect Plasma: any off‑chain plasma protocol must either incur large exit sizes or allow non‑uniquely attributable mass exits. Under…
Talk on
Majorizing Measures for the OptimizerSander Borst, Daniel Dadush, Neil Olver, Makrand Sinha33m
Proves a constructive version of the majorizing measures duality: a convex program for the Gaussian chaining functional and a simplified dual are within a…
Talk
Randomness and Fairness in Two-Sided Matching with Limited InterviewsHedyeh Beyhaghi, Eva Tardos20m
Model has N applicants and N positions, each side draws iid uniform preferences, applicants send K applications, positions interview K' and then…
Talk
Counterexamples to the Low-Degree ConjectureJustin Holmgren, Alexander S Wein19m
Refutes Hopkins’ low-degree conjecture by building distributions where degree-O(log n) polynomials provably fail to distinguish IID Gaussian Q from structured…
Talk on
High-entropy dual functions and locally decodable codesFarrokh Labib, Jop Briet34m
Shows a negative result: for infinitely many primes P there exists a K(P) and explicit families of order-k dual functions over F_p^n that cannot be…
Talk on
Buying Data Over Time: Approximately Optimal Strategies for Dynamic Data-Driven DecisionsNicole Immorlica, Ian Kash, Brendan Lucier30m
Model where a one-dimensional hidden state follows a drifting Gaussian state and noisy samples cost a per-sample fee plus optional fixed setup cost…
Talk on
Batching and Optimal Multi-stage Bipartite AllocationsYiding Feng, Rad Niazadeh24m
A polynomial-time algorithm achieves the optimal competitive ratio for K-stage bipartite allocation, namely gamma_K = 1 - (1-1/K)^K (tending to 1-1/e), and a…
Talk
Learning and Strongly Truthful Multi-task Setting Peer Prediction: A Variational ApproachGrant Schoenebeck, Fang-Yi Yu27m
Defines a variational estimator of phi‑mutual information and uses it to build a multi‑task peer‑prediction payment: learn an ideal scoring rule K on withheld…
Talk on
Distributed load balancing: a new framework and improved guaranteesAllen Liu, Sara Ahmadian, Binghui Peng, Morteza Zadimoghaddam19m
Distributed algorithm balances loads on a bipartite source–worker graph in the CONGEST model, producing a 1+ε approximation for general convex objectives by…
Talk on
Erasure-Resilient Sublinear-Time Graph AlgorithmsAmit Levi, Ramesh Krishnan S. Pallavoor, Sofya Raskhodnikova, Nithin Varma25m
Defines queries on an alpha-erased graph (adjacency lists with adversarial missing entries) and measures performance over all completions; degree and neighbor…
Talk on
How to Sell Information Optimally: an Algorithmic StudyYang Cai, Grigoris Velegkas22m
Provides an algorithmic characterization: when type distribution and payoff matrix are explicit the revenue-optimal menu is computable via a polynomial-size…
Talk
Computation Over the Noisy Broadcast Channel with Malicious PartiesKlim Efremenko, Gillat Kol, Dmitry Paramonov, Raghuvansh R. Saxena27m
Shows a protocol that computes bit‑exchange over a noisy broadcast channel tolerating up to half malicious parties while cutting per‑player communication from…
Talk on
~/search
type to search 6,249 conferences, 138,987 speakers and 185,172 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.