2023 Fiscal Year Research-status Report
Indexing Massive Datasets with Algorithmic Engineered Compression Techniques on Modern Computer Architectures
| Project/Area Number |
21K17701
|
|---|
| Research Institution | University of Yamanashi |
|---|
Principal Investigator |
Koeppl Dominik 山梨大学, 大学院総合研究部, 特任准教授 (50897395)
|
|---|
| Project Period (FY) |
2021-04-01 – 2025-03-31
|
| Keywords | compressed indexes / string subsequences / NP-hard problems / straight line programs / collage systems / block trees / parameterized BWT / pattern matching |
|---|
| Outline of Annual Research Achievements |
Following the research plan outlined for fiscal year 2023, our primary focus was on extending string regularities from substrings to subsequences, exploring NP-hard problems associated with strings, and refining compressed indexing data structures.
In the first thematic area, for computing the longest Lyndon subsequence, we achieved space and time bounds superior to those presented at IWOCA in 2022. Furthermore, we demonstrated methodologies for computing the longest bordered and periodic subsequences. This involved using novel tools to compute the longest common subsequences between all prefixes and suffixes of a text, which facilitated the computation of longest bordered or periodic subsequences. Asides, for the longest bordered subsequences, we established a conditional lower bound aligning with our quadratic running time.
Subsequently, we delved into studying common NP-hard problems with strings as inputs, leveraging answer set programming solvers. Additionally, we proved the NP-hardness of finding the smallest run-length compressed straight-line programs (RLSLPs) for unbounded alphabet sizes. We could adapt this proof to finding the smallest collage system. Additionally, we devised a MAX-SAT encoding for computing the smallest RLSLP.
In the final thematic area, we made advancements in the construction, practically for block trees and theoretically for the parameterized Burrows-Wheeler transform.
For the latter, we also demonstrated that this transform can be adapted for circular pattern matching by changing the encoding.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
We conducted the research for the fiscal year 2023 as planned,
and could complete most of our planned research at the end of the grant lifespan in the fiscal year 2023.
|
| Strategy for Future Research Activity |
As the grant's term ended in fiscal year 2023, we are now in the process of preparing to apply for a new grant for fiscal year 2025, based on the fact that this research has unveiled new paths for further exploration within the realm of string regularities and compressed indexes, igniting our enthusiasm to pursue these paths in the forthcoming years.
While our main attention has been set to text indexing data structures for classic pattern matching, the exploration of extended pattern matching queries remains largely undone. In response, we aim to expand upon several concepts discovered during our recent research, combining them with cutting-edge indexing techniques tailored for classic pattern matching. We anticipate that these innovative indexing methodologies will find practical applications in scenarios where conventional pattern matching proves too restrictive, necessitating more adaptable matching criteria.
|
