TU WIen HauptgebΓ€ude
Hallo Bluesky-Community!
Die TU Wien teilt ab sofort spannende News zu Forschung und Wissenschaft hier auf Bluesky.
Bild c_Matthias Heisler
@mlanzinger.bsky.social
Assistant Professor TU Wien, previously Uni of Oxford | Research in #databasetheory, #AI & #GNNs
TU WIen HauptgebΓ€ude
Hallo Bluesky-Community!
Die TU Wien teilt ab sofort spannende News zu Forschung und Wissenschaft hier auf Bluesky.
Bild c_Matthias Heisler
New paper: Simulating Time With Square-Root Space
people.csail.mit.edu/rrw/time-vs-...
It's still hard for me to believe it myself, but I seem to have shown that TIME[t] is contained in SPACE[sqrt{t log t}].
To appear in STOC. Comments are very welcome!
The (soft) deadline for this position is on Friday.
17.02.2025 07:16 β π 5 π 5 π¬ 0 π 0Anyone want to start a betting pool on the % of #iclr meta-reviews that are fully LLM generated?
22.01.2025 17:32 β π 1 π 0 π¬ 0 π 05/5
Finally, a prototype implementation demonstrates the practical impact on query evaluation. Early experiments show potential for large performance gains on difficult queries in standard database systems (e.g., PostgreSQL, SparkSQL)
Check it out: arxiv.org/pdf/2412.11669
4/
β’ SHW can achieve smaller widths than HW while remaining tractable to compute, boosting efficiency.
β’ SHW induces a hierarchy of width measures that equal generalised hypertree width in the limit.
3/ π Our contribution:
We propose Soft Hypertree Width (SHW) β a relaxed version of hypertree width that provides additional algorithmic flexibility, which allows for efficiently incorporating preferences and constraints into query decompositions.
2/ The problem: Theoretical performance of queries in databases depends on the "width" of decompositions of the query. The smaller the βwidthβ, the more efficient evaluation is. But real-world performance isnβt just about the smallest widthβ¦ constraints & preferences matter too!
20.12.2024 08:38 β π 0 π 0 π¬ 1 π 01/ π’ Excited to share our new paper: βSoft and Constrained Hypertree Widthβ β a step forward in optimizing database queries by introducing a flexible, and effective hypergraph decomposition framework. π§΅ arxiv.org/pdf/2412.11669
20.12.2024 08:38 β π 3 π 0 π¬ 1 π 0Open Tenure Track Positions at the University of Vienna! π¨βπ¬ π©βπ¬
#univie is currently accepting applications for several
#tenuretrack positions across a range of disciplines. π§ͺ π
Work with us to find answers to the questions for tomorrow. π ‡ #academicsky
jobs.univie.ac.at/go/Tenure-Tr...
Happy to share my recent interview with TU Wien Informatics, where I discuss algorithms, social network analysis, and the beauty of theoretical computer science.
Check it out here: informatics.tuwien.ac.at/news/2713
#TCS #Algorithms #SocialNetworks
The PCP theorem, a jewel of theoretical computer science, establishes that any NP statement can be assessed by a randomized verifier who only checks a vanishing fraction of the proof (indeed, a constant # of characters!)
This has had incredible impact, most notably on how ML reviews are conducted
Also more recently arxiv.org/abs/2410.18676 :)
25.11.2024 14:18 β π 1 π 0 π¬ 0 π 0Emily and I have been working towards this moment for years. Two big papers just to set up one obscure reference to 90s hip-hop. Feeling proud to have finally achieved it.
25.11.2024 11:31 β π 25 π 1 π¬ 2 π 0