Please enable JavaScript.
Coggle requires JavaScript to display documents.
Exploiting Parallelism in Document Similarity Tasks with Application (3.…
Exploiting Parallelism in Document Similarity Tasks with Application
1.Introduction
Work Proposal
Hypothesis
Proposed Implementations
Motivation
Contributions
Work Organization
2. Parallel Computing
Parallel Programming Models
BSP
Parallel Architectures
Taxonomy of computers
CPU Multicore
GPU Manycore
Summary
3. Text Representation and
Document Similarity Search
Bag of words
Word/Document Embedding
Document Similairity Search
Related works
Summary of proposals
5. Parallel approach to accelerate word embedding generation with fine-grain parallelism
The original W2V in a multicore architecture
Parallel Solution
Application (automated program repair)
Challenges for word embbedings generation
w2v
glove
7. Conclusions
Summary of contributions
Published papers
Manuscripts in review process
Discussion
Limitations of our proposal
Future works
4. Parallel approach to Document Similarity Search and Top-K Applications
kNN
Challenges
Related works
GT-KNN
GKNN
Others
Proposed work
Data Indexing
K Nearest Neighbors Search
Threshold-based FIltering
FaSST-kNN
Multi-query KNN Search
SFaSST-kNN
8. References
6. Experiments and Results
kNN
Embeddings