msuzgun.bsky.social
JD-PhD Candidate in Computer Science at Stanford
👤 Joined on 22 November 2024
Followers
Following
Total Posts
@msuzgun.bsky.socialstring2string is super fun and useful! Compare and visualize texts and sequences in your browser. Accepted as an EMNLP demo. Great job @msuzgun.bsky.social!
Show Post10/ We hope string2string Studio supports your research & practical work. We warmly welcome feedback, feature requests, and contributions. And we'd be delighted to exhibit projects & findings developed with it. - Paper: arxiv.org/abs/2608.03984 - Website: string2string.org
Show Post9/ We validate against independent implementations. Under declared settings, core algorithms and metrics agree exactly with their references; our scoped blastn path closely matches NCBI BLAST+. The Studio does not replace specialist tools at every scale.
Show Post8/ Our internal benchmarks show that the engine runs about 100-2,500× faster than our Python predecessor and that it is competitive with fast, native alignment libraries on representative real-world settings.
Show Post7/ The interface is built for exploration: change tokenization, scoring rules, gap penalties, or metric settings and see the result update in place. The curated "Showcase" and "Learn" modules turn canonical examples into editable demos.
Show Post