Suhail

Unconditional Lower-Bound Barrier for AC0-Natural Proofs Against Constant-Depth Circuits

This paper establishes an unconditional lower-bound barrier for AC0-natural proofs against constant-depth circuits. While natural proofs were previously shown to have limitations under cryptographic assumptions, this work provides an unconditional result by adapting the Trevisan-Xue pseudorandom generator. The derived barrier, $2^{n^{7/(d-5)}}$, is quantitatively close to the frontier achieved by the Switching Lemma ($2^{n^{1/(d-1)}}$), with a notable self-referential aspect where the proof utilizes the Switching Lemma itself to demonstrate this limitation.

Front-to-Attractors: A Computationally Efficient Heuristic for Bidirectional Search

This paper introduces Front-to-Attractors (F2A), a novel heuristic for bidirectional search that improves upon the Front-to-Front (F2F) heuristic. F2A significantly reduces computational overhead by estimating distances to a small set of dynamically maintained attractors rather than the entire opposite search frontier. This approach preserves the informativeness of F2F heuristics while drastically cutting down on pairwise evaluations and yielding fewer node expansions compared to Front-to-End (F2E) heuristics.

A Natural $\gtrsim 100\times$ Telescope: Discovery of the Strongly Lensed Type II SN 2025mkn at $z=1.37$

reward-lens: A Mechanistic Interpretability Library for Reward Models

Graph-State Circuit Blocks control Entanglement and Scrambling Velocities