Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2008.12825

Is the space complexity of planted clique recovery the same as that of detection?

Jay Mardia
View original source ↗Read source document ↗Export metadata ↓

Abstract

We study the planted clique problem in which a clique of size k is planted in an Erdős-Rényi graph G(n, 1/2), and one is interested in either detecting or recovering this planted clique. This problem is interesting because it is widely believed to show a statistical-computational gap at clique size k=sqrt{n}, and has emerged as the prototypical problem with such a gap from which average-case hardness of other statistical problems can be deduced. It also displays a tight computational connection between the detection and recovery variants, unlike other problems of a similar nature. This wide investigation into the computational complexity of the planted clique problem has, however, mostly focused on its time complexity. In this work, we ask- Do the statistical-computational phenomena that make the planted clique an interesting problem also hold when we use `space efficiency' as our notion of computational efficiency? It is relatively easy to show that a positive answer to this question depends on the existence of a O(log n) space algorithm that can recover planted cliques of size k = Omega(sqrt{n}). Our main result comes very close to designing such an algorithm. We show that for k=Omega(sqrt{n}), the recovery problem can be solved in O((log*{n}-log*{k/sqrt{n}}) log n) bits of space. 1. If k = omega(sqrt{n}log^{(l)}n) for any constant integer l > 0, the space usage is O(log n) bits. 2.If k = Theta(sqrt{n}), the space usage is O(log*{n} log n) bits. Our result suggests that there does exist an O(log n) space algorithm to recover cliques of size k = Omega(sqrt{n}), since we come very close to achieving such parameters. This provides evidence that the statistical-computational phenomena that (conjecturally) hold for planted clique time complexity also (conjecturally) hold for space complexity.

Explore related subjects

cs.CCcs.DScs.ITcs.LGmath.IT

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Jay Mardia. 2020-11-23. Is the space complexity of planted clique recovery the same as that of detection?. https://arxiv.org/abs/2008.12825

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Bit-counting complexity classes

We define bit-counting complexity classes whose membership depends on the binary profile of the number of accepting paths of non-deterministic polynomial time Turing machines. We study the relationship between this new family of complexity classes and the classical complexity classes. We prove that the complexity class ${\bf PP}$ is contained in our comparison based bit-counting complexity classes ${\bf B_{|0|=|1|}P}$, ${\bf B_{|0|<|1|}P}$ and ${\bf B_{|0|>|1|}P}$. We then show that the comparison based bit-counting complexity classes and the complexity class ${\bf PP}$ are Turing equivalent, that is ${\bf P}^{\bf PP} = {\bf P}^{{\bf B_{|0|=|1|}P}}={\bf P}^{{\bf B_{|0|>|1|}P}}={\bf P}^{{\bf B_{|0|<|1|}P}}$. We then prove that the complexity classes ${\bf NP}$ and ${\bf CoNP}$ are contained in both of our parity based bit-counting complexity classes ${\bf B_{|0| \oplus}P}$ and ${\bf B_{|1| \oplus}P}$. We also show that the Turing closures of the parity based bit-counting complexity classes coincide, that is ${\bf P}^{{\bf B_{|0|\oplus}P}}={\bf P}^{{\bf B_{|1|\oplus}P}}$. We do this by proving that when either parity based bit-counting complexity class is provided as an oracle for a polynomial time Turing machine, then it can simulate the other one, that is ${\bf B_{|1| \oplus}P}\subseteq {\bf P}^{{\bf B_{|0| \oplus}P}}$ and ${\bf B_{|0| \oplus}P}\subseteq {\bf P}^{{\bf B_{|1| \oplus}P}}$.

cs.CC↗
arXiv2026

Self-Referential $K$-SAT and the Finite Analogue of Gödel's Incompleteness Theorem

Self-reference and solution independence are central to hard combinatorial instances. We ask whether Boolean \(K\)-SAT can exhibit both, giving a finite propositional analogue of Gödel's incompleteness theorems. Solution independence is formalized via factorial moments of the satisfying-assignment count. Constant-width random \(K\)-SAT fails: overlapping assignments create correlations and exponential second moment. We use a random CNF ensemble with logarithmic width \(K=O(\log N)\) at the subcube-covering threshold \(M=Θ(N^{2+\varepsilon})\). There it converges to Poisson, so unsatisfiable and uniquely satisfiable formulas coexist. Using the unique solution, a single-clause replacement yields a SAT/UNSAT pair sharing the same unsigned incidence graph. We prove structural irreducibility: every local subinstance of size at most \(N^c\), \(0<c<1\), has identical local views, so no deterministic or bounded-error clause-query evaluator at that scale can distinguish unique satisfiability from unsatisfiability. This is a finite self-referential construction, not itself a time-complexity lower bound. We quantify the local--global gap: any transcript of \(t=N^{1-δ}\) queries leaves \(N-o(N)\) bits of witness entropy. Expansion preservation plus size--width, size--degree, and pseudoexpectation trade-offs gives linear Resolution width, linear PC/PCR and SOS degree, and exponential proof size for the unsatisfiable companions; analogous bounds hold for semantic Cutting Planes and restricted Positivstellensatz. These bounds are uniform over support-preserving signings, even after observing the unique solution. These \(2^{Ω(N)}\) proof-size lower bounds are consistent with SETH, suggesting SETH is a finite projection of Gödel incompleteness onto resource-bounded computation. The hardness is quantum-invariant and limits local statistical learning.

cs.CC↗
arXiv2026

SC Derandomization for Regular ROBPs and Models Beyond BPL

We study SC derandomizations for regular read-once branching programs (ROBPs) and computation models beyond BPL. For regular ROBPs with length $n$, width $w$, and multiple accept nodes, we attain three results. 1. When $n \le w$, we show an SC derandomization with space $O(\log^2 n+\log w)$ and error $1/\text{poly}(nw)$. 2. When $n \ge w$, we show an SC derandomization with space $O(\log n \log w)$ and error $1/\text{poly}(w)$. 3. When $w=O(\log n)$, we show an optimal $O(\log n)$ space derandomization with error $1/\text{poly}(w)$. We further show that two super sets of BPL can be computed in SC. 1. For probabilistic logspace TMs with a two-way access random tape, we show that it can be approximated in SC if each entry of the random tape is accessed for at most a constant number of times. 2. For probabilistic logspace TMs with a polynomial size stack, i.e. probabilistic logspace Auxiliary Push-down Machines (AuxPDMs), we show that it can be approximated in SC if the timings of push/pop/idle stack operations do not depend on the randomness. The first model is the read-multiplicity model considered by Impagliazzo, Nisan, Wigderson (STOC'94), in which they show that their INW generator can fool such computations. For the second model, we indicate that it contains candidate languages separating BQL from BPL considered by Apers and Edenhofer (CCC'25).

cs.CC↗