Topic Similarity in Information Retrieval
Examples and Experience of NLP Centre and LEMMA Projects
Petr Sojka
Laboratory of Electronic and Multimedia Applications and
Natural Language Processing Centre, Faculty of Informatics
Masaryk University, Brno, Czech Republic
sojka@fi.muni.cz
PV211 Intro to Information Retrieval: LDA
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 1 / 17
Coping with Information Overload by Filtering of Big Data
Life is searching: group similar and narrow focus of search in [your]
Big Data.
Similarity types: from plagiarism (similarity on n-grams, narrative
similarity, evolved into http://theses.cz) to thematic, topical
similarity.
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 2 / 17
Prehistoric Example: Project Ottův Slovník naučný, 1998
Levels of content processing: strings → words and collocations →
semantics (word meaning) → information (knowledge).
Grabbing the essence (content) of documents: topical modelling.
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 3 / 17
Topical Similarity in Digital Mathematics Library
2005, GVP, Radim Řehůřek and Jan Pomikálek
2006, gensim, different machine learning methods as Random
Projections, TFIDF word weighting, Latent Semantic
Indexing/Analysis, Latent Dirichlet Allocation
50,000+ full-texts on http://dml.cz
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 4 / 17
Leading Edge Example: Automated Meaning Picking from
Texts
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 5 / 17
Probabilistic Topical Modelling: Latent Dirichlet Allocation
topic: weighted list of words
document: weighted list of topics
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 6 / 17
Topical Modelling: Latent Dirichlet Allocation II
all topics computed automatically from document corpora
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 7 / 17
Content Similarity Results in EuDML
Within European Digital Mathematics Library, EuDML, project EU
CIP-ICT-PSP we have developed and delivered technology for
similarity (gensim), document conversions (Braille) and accessibility
(math OCR), NLP content normalization (Mathml2text).
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 8 / 17
Math Search Interface EuDML
Demo of math search in EuDML
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 9 / 17
Digital Library Service Architecture and Workflow (EuDML)
Document engineering and workflows including [Math] OCR.
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 10 / 17
Digital Library Service Architecture and Workflow
(DML-CZ)
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 11 / 17
Data Visualization and Representation
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 12 / 17
Award Winning Topic Similarity Framework gensim
Semantic similarity indexing and search of big (continuous
stream of) data. Client (search) and server (indexing)
architecture.
Developed by NLPlab PG student Radim Řehůřek (awarded in
Česká hlava competition in 2011).
Leading edge machine learning methods implemented.
Used in 60+ local, EU or worldwide projects, 260+ citations.
Typical deployment and fine-tuning scenario: expressing data
as words (features) → configuration of topic modelling of
features → setting of gensim methods and tuning parameters
→ usage in an application with proper visualization interface.
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 13 / 17
Teaching Laboratory build with Constructivism Principles
most work done by students themselves with agile techniques,
XP
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 14 / 17
Conclusions and Mutual Research Interests
similarity by topical modelling,
document filtering and visualization
semantic, meaning computations and modelling of natural
language texts (natural NLP)
personal research interests: random walking for
disambiguation, math (tree) indexing and similarity
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 15 / 17
That’s it!
Yes, we can!
Credits: Jiří Franek (illustrations)
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 16 / 17
Links
NLP Centre: http://nlp.fi.muni.cz/
Topical modelling: https://mir.fi.muni.cz/gensim/
Math Information Retrieval: https://mir.fi.muni.cz
DML-CZ project: http://dml.cz, http://project.dml.cz
EuDML project: http://eudml.cz,
http://project.eudml.cz
LEMMA: http://www.fi.muni.cz/lemma/
Petr Sojka: Topic Similarity in Information Retrieval May 3rd, 2016 17 / 17