Tw:Brahms Medha Presentation 0918

From Semantic Portal Wiki

Presentation given at CSCI 6966 Advanced Semantic Web (Fall 2008) - Lesson 4

• Speaker: Medha Atre
• Title: BRAHMS: A WorkBench RDF Store And High Performance Memory System for Semantic Association Discovery
• Authors: Maciej Janik, Krys Kochut
• Conference: ISWC2005
• URL: http://lsdis.cs.uga.edu/library/download/JK05-ISWC2005.pdf
• Date of Presentation: 2008/09/18

Questions

ID	Question	Name	Answer
Brahms GregoryToddWilliams Question1	Is the Brahms system meant only to answer association queries, or to answer association queries in conjunction with more general graph patterns? The comparison of Brahms to three systems implemented, in part, for general query answering (with the expressive power of SPARQL) seems a bit unfair. Brahms is implemented in a way so as to speed association queries while ignoring features of more general query languages. In contrast, the three systems it is compared to use techniques that speed SPARQL query execution, but dramatically slow down association queries. While this point of difference is legitimately made, a fairer comparison might have been made by, for example, removing the default compound indexes in Redland (which are never used in the association discover algorithms) at the same time as adding the new association-specific index. Such removal would have a direct affect on the memory usage tests presented in section 6.2 in which the "Redland, IDX" variant ran out of memory.	Gregory Todd Williams
Brahms Medha Presentation 0918 Jesse Weaver	Have the authors done any subsequent research using regular paths as indicated in the Future Work section? This paper seems to make more feasible association discovery, but queries like "find any connection between A and B" don't seem to take full advantage of available semantics. (For example, just because two people recommend the same wine for the same dish doesn't really mean they're associated in a very meaningful way.) It seems like using regular paths may help constrain the query to something more meaningful than just any connection.	Jesse Weaver
Brahms Medha Presentation 0918 Joshua Taylor Question 1	The authors seem to have built an RDF store for the sole purpose of discovering semantic associations (paths connecting resources), and their store does seem to be effective. However, their motivation stems from the difficulties that other, more general RDF stores had in achieving this task. Yet no pre-processing seems to have been performed on the RDF graphs. In their Future Work section, the authors state their intention "to experiment with a variety of semantic association discovery algorithsm, utilizing a language for defining regular paths … . The regular expressions defined over the RDF resources and types … will enable us to define the association paths of interesting patterns and significantly restrict the search space of the semantic association discovery." If only certain bits of the graph are interesting, and BRAHMs isn't a general-purpose RDF store, why not just throw out uninteresting triples? And if this approach can be taken, why not ease the burden of the other systems by only asking them to store the interesting triples. Perhaps graphs might then be small enough for them to store.	Joshua A. Taylor
Brahms Medha Presentation 0918 Joshua Taylor Question 2	The authors do not seem to cite the particular depth-first search that they are using, but it does not seem to compare well with their bi-directional breadth first search. Is there any indication of whether they are using a straightforward depth first search versus an iterative deepening search (as discussed by, e.g., Russell and Norvig in "Artificial Intelligence: A Modern Approach")? On a related note, the authors mention in their "Future Work" section that they plan on experimenting with languages for expressing regular paths. This suggests that they recognize some predicates and resources as more interesting than others. They have developed an efficient RDF store for performing graph based search algorithms, and seem to have a notion of cost, or benefit, on the edges and nodes of the graph—why do they not mention using heuristic algorithms?	Joshua A. Taylor
Brahms Ques Ankesh	* @Pg 7-"executing a bi-directional breadth-first search (bi-BFS) utilizing a trie representation of the search structures in order to find semantic associations... I couldn't understand the search structure outlined here @Pg 10- Memory usage for data-sets of sizes 14, 255, 14, 556 are 20, 270, 10, 501 MBs respectively. In 2 cases there have been significant %age drop in storage sizes and in other 2 there have been increase in storage sizes. How can this be explained? (number of resources vs. number of triples for each of them?) Casual Question- The motivation is drawn from examples in Anti-money Laundering, Threat Assessment and Risk Assessment. However, the system has been tested only for Insider Threat Project, whose expected data set size is small. Aren't above data sets too huge, to keep memory based associations feasible?	Ankesh Khandelwal
Janik2005brahms question 1 by lebo	How many implementations of the depth-first search and bi-directional breadth-first exhaustive search algorithms were used in the evaluation? Three languages were used (C++ (BRAHMS), C (Redland), and Java (Jena and Sesame)) among the four triple stores. If multiple implementations were used, what assurances were made that they performed similarly before incorporating the triple store?	Tim Lebo
Janik2005brahms question 2 by lebo	Did the authors repeat the load and execution tests to demonstrate reproducability? How much variability in the results could we expect?	Tim Lebo

Attendees

Tim Lebo