Zitan Chen

Xiao Cao, Zitan Chen, Canyu Le et al.

Chapter generation becomes practical technique for online videos nowadays. The chapter breakpoints enable users to quickly find the parts they want and get the summative annotations. However, there is no public method and dataset for this task. To facilitate the research along this direction, we introduce a new dataset called Chapter-Gen, which consists of approximately 10k user-generated videos with annotated chapter information. Our data collection procedure is fast, scalable and does not require any additional manual annotation. On top of this dataset, we design an effective baseline specificlly for video chapters generation task. which captures two aspects of a video,including visual dynamics and narration text. It disentangles local and global video features for localization and title generation respectively. To parse the long video efficiently, a skip sliding window mechanism is designed to localize potential chapters. And a cross attention multi-modal fusion module is developed to aggregate local features for title generation. Our experiments demonstrate that the proposed framework achieves superior results over existing methods which illustrate that the method design for similar task cannot be transfered directly even after fine-tuning. Code and dataset are available at https://github.com/czt117/MVCG.

Huiling Xie, Zitan Chen

We study error-correcting codes for an adversarial nanopore channel, where a $q$-ary string is first transformed by an inter-symbol interference channel with window size $\ell$ into a sequence of overlapping $\ell$-mers, and an adversary then corrupts this $\ell$-mer sequence by introducing at most $t$ edits. For the deletion-only nanopore channel, we show that the optimal redundancy of $t$-deletion-correcting codes of length $n$ lies between $t\log_q n+Ω(1)$ and $2t\log_q n-\log_q\log_2 n+O(1)$. We then give two explicit deletion-correcting constructions in the regime $t\leq \min\{(\ell-1)/2,(\ell+2)/3\}$. The first construction relies on generalized Reed-Solomon codes and has redundancy $2t\log_q n+Θ(\log\log n)$. The second is based on Sidon sets (or rather $B_t$ sequences) and has redundancy $t\log_q n+Θ(\log\log n)$, matching the lower bound to first order. We further extend the $B_t$-based approach to the edit channel, allowing insertions, deletions, and substitutions of $\ell$-mers. In the regime $t\leq \min\{(\ell-1)/4,(\ell+2)/6\}$, this gives explicit $t$-edit-correcting codes with redundancy $t\log_q n+Θ(\log\log n)$, which is first-order optimal.