arXiv subjects

Jai Aslam

Publications and source records attributed to Jai Aslam.

4 recordsLinked to original sources

arXiv2026

Advancing Reliable Synthetic Video Detection: Insights from the SAFE Challenge

The proliferation of generative video technologies has intensified the need for reliable methods to detect and characterize synthetic media. To address this challenge, we organized the \href{https://safe-video-2025.dsri.org}{SAFE: Synthetic Video Detection Challenge}, co-located with the \textit{Authenticity and Provenance in the Age of Generative AI (APAI) Workshop }at ICCV 2025. The competition invited participants to develop and evaluate algorithms capable of distinguishing real from synthetic videos under fully blind evaluation conditions with over 600 submissions from 12 teams over a 90 day span. Hosted on the Hugging Face platform, the challenge comprised two primary tasks: (1) detection of synthetic video content generated by diverse state-of-the-art models, and (2) detection of synthetic content following common post-processing operations such as resizing, re-compression, motion blur and others. The challenge data consisted of 13 modern high quality synthetic video models with generated content matched to real videos from 21 diverse and challenge sources, all adding up to 20 hours of 6,000 video samples. This paper describes the challenge design, dataset construction, evaluation methodology, and outcomes, offering insights into the generalization and robustness of contemporary synthetic video detection methods. Our findings highlight measurable progress in cross-generator generalization but also persistent vulnerabilities to post-processing artifacts. https://safe-video-2025.dsri.org

cs.CV↗

arXiv2025

Synthetic Audio Forensics Evaluation (SAFE) Challenge

The increasing realism of synthetic speech generated by advanced text-to-speech (TTS) models, coupled with post-processing and laundering techniques, presents a significant challenge for audio forensic detection. In this paper, we introduce the SAFE (Synthetic Audio Forensics Evaluation) Challenge, a fully blind evaluation framework designed to benchmark detection models across progressively harder scenarios: raw synthetic speech, processed audio (e.g., compression, resampling), and laundered audio intended to evade forensic analysis. The SAFE challenge consisted of a total of 90 hours of audio and 21,000 audio samples split across 21 different real sources and 17 different TTS models and 3 tasks. We present the challenge, evaluation design and tasks, dataset details, and initial insights into the strengths and limitations of current approaches, offering a foundation for advancing synthetic audio detection research. More information is available at \href{https://stresearch.github.io/SAFE/}{https://stresearch.github.io/SAFE/}.

cs.SD↗

arXiv2018

Splitting loops and necklaces: Variants of the square peg problem

Toeplitz conjectured that any simple planar loop inscribes a square. Here we prove variants of Toeplitz' square peg problem. We prove Hadwiger's 1971 conjecture that any simple loop in $3$-space inscribes a parallelogram. We show that any simple planar loop inscribes sufficiently many rectangles that their vertices are dense in the loop (independently due to Schwartz). If the loop is rectifiable, there is a rectangle that cuts the loop into four pieces that can be rearranged to form two loops of equal length. A rectifiable loop in $d$-space can be cut into $(r-1)(d+1)+1$ pieces that can be rearranged by translations to form $r$ loops of equal length. We relate our results to fair divisions of necklaces in the sense of Alon and to Tverberg-type results. This provides a new approach and a common framework to obtain variants of Toeplitz' square peg problem for the class of all continuous curves.

math.MG↗

arXiv2017

On the generalized Erdős--Kneser conjecture: proofs and reductions

Alon, Frankl, and Lovász proved a conjecture of Erdős that one needs at least $\lceil \frac{n-r(k-1)}{r-1} \rceil$ colors to color the $k$-subsets of $\{1, \dots, n\}$ such that any $r$ of the $k$-subsets that have the same color are not pairwise disjoint. A generalization of this problem where one requires $s$-wise instead of pairwise intersections was considered by Sarkaria. He claimed a proof of a generalized Erdős--Kneser conjecture establishing a lower bound for the number of colors that reduces to Erdős' original conjecture for ${s = 2}$. Lange and Ziegler pointed out that his proof fails whenever $r$ is not a prime. Here we establish this generalized Erdős--Kneser conjecture for every $r$, as long as $s$ is not too close to $r$. Our result encompasses earlier results but is significantly more general. We discuss relations of our results to conjectures of Ziegler and of Abyazi Sani and Alishahi, and prove the latter in several cases.

math.CO↗