Medical Computer Science

Curriculum Vitae Youngho Kim (abpavw@r.postjobfree.com)

Youngho Kim

Center for Intelligent Information Retrieval abpavw@r.postjobfree.com

*** ********* ***** *****://www.cs.umass.edu/~yhkim/

University of Massachusetts, Amherst Cell: +1-413-***-****

MA, 01003, USA. Office: +1-413-***-****

Research Interest Information Retrieval, Information Extraction, and Natural Language Processing.

University of Massachusetts Amherst

Education

Ph.D. student in Computer Science, 2009 ~ present.

Academic Advisor: Professor W. Bruce Croft

Korea Advanced Institute of Science and Technology (KAIST)

M.S. in Information and Communications Engineering, 2009.

Academic Advisor: Professor Sung-Hyon Myaeng

Thesis: A bootstrapping technique for acquiring domain -specific feature in

sentiment analysis

INHA University

B.S. in Computer Science & Engineering, 2006 .

Publication Refereed

Youngho Kim, Jangwon Seo, and W. Bruce Croft. Automatic Boolean Query

Suggestion for Professional Search In Proc. of the 34th Int l ACM SIGIR

Conference on Research and Development in Information (SIGIR 11), 2011.

Youngho Kim, Yoonjung Choi, and Sung-Hyon Myaeng. Generating Domain-

specific Clues using News Corpus for Sentiment Classification In Proc. of the 4th

Int l AAAI Conference on Weblogs and Social Media (ICWSM 10), 2010. (poster)

Youngho Kim, Jihee Ryu, and Sung-Hyon Myaeng. A Patent Retrieval Method

using Semantic Annotations In Proc. of Int l Conference on Knowledge Discovery

and Information Retrieval (KDIR 09), 2009.

Yoonjung Choi, Youngho Kim, and Sung-Hyon Myaeng. Domain-specific

Sentiment Analysis using Contextual Feature Generation In Proc. of the 1st Int l

CIKM Workshop on Topic-Sentiment Analysis for Mass Opinion Measurement,

2009.

Youngho Kim, Yingshi Tian, Yoonjae Jung, Jihee Ryu, and Sung-Hyon Myaeng.

Automatic Discovery of Technology Trends from Patent Text In Proc. of the

24th Annual ACM Symposium on Applied Computing (SAC 09), 2009.

Youngho Kim, Yuchul Jung, and Sung-Hyon Myaeng. An Opinion Analysis

System Using Domain-Specific Lexical Knowledge. In Proc. of the 4th Asia

Information Retrieval Symposium (AIRS 08), 2008. (poster)

Yuchul Jung, Joo-Young Lee, Youngho Kim, Jaehyun Park, Sung-Hyon Myaeng,

and Hae-Chang Rim. Building a Large-Scale Commonsense Knowledge Base by

Converting an Existing one in a Different Language In Proc. of the 8th Int l

Conference on Intelligent Text Processing and Computational Linguistics

(CICLing 07), 2007.

Curriculum Vitae Youngho Kim (abpavw@r.postjobfree.com)

Un-refereed

Youngho Kim, Seongchan Kim, and Sung-Hyon Myaeng. Extracting Topic-

related Opinions and their Targets in NTCIR-7 In Proc. of the 7th NTCIR

Evaluation Workshop, 2008.

Youngho Kim and Sung-Hyon Myaeng. Opinion Analysis based on Lexical

Clues and their Expansion In Proc. of the 6th NTCIR Evaluation Workshop, 2007.

(selected for oral presentation)

Microsoft Research Asia

Work Experience

Research Intern, Aug. 2007 ~ Feb. 2008.

Worked in the Natural Language Computing group leaded by Dr. Ming Zhou.

Research

Effective Query Generation for Medical Search (Oct. 2011 ~ present)

Experience

Given a new medical statement, finding relevant web pages to provide information to

casual users (e.g., patients).

Generating effective queries from medical statements (incl. illness history,

symptoms, and objective signs).

External resource-based approaches using medical ontology, Medline pages, and

medical forum pages.

Devising the method to bridge language gaps between medical documents and plain

English texts.

Providing a summary of generated queries for casual users.

Effective Query Generation for Academic Searches (Feb. 2011 ~ Sept. 2011)

Papers are in preparation.

Given a research description, finding relevant academic articles (literature search).

Generating concept queries to contain Key Concepts (effective phrase to retrieve

relevant articles) and Related Concepts (co-occurring phrases with the key

concepts, which provide context information for the key concepts).

Applying the label propagation to find effective phrases by propagating retrieval

effectiveness from a baseline query.

Devising concept-query specific retrieval features.

Integrating concept-query features with state-of-the-art features (e.g., citation

network, author behaviors, recency of academic articles, and etc.).

Boolean Query Suggestion for Professional Search (Jan. 2010 ~ Jan. 2011)

Presented at SIGIR 11.

Motivated from the necessity of Boolean operators in Professional Search (domain-

specific searches performed by information professionals, e.g., patent search).

Boolean queries containing conjunction and negation are strongly preferred by

professional users.

Generating Boolean queries using Decision Trees learned from pseudo-labeled

documents by a baseline system (incl. long keyword query)

Utilizing the query quality predictors to rank generated Boolean queries and suggest

top-k queries as candidates.

Domain-specific Sentiment Analysis (Aug. 2008 ~ June 2009)

Curriculum Vitae Youngho Kim (abpavw@r.postjobfree.com)

Research for Master Thesis.

Presented at ICWSM 10.

Developed a bootstrapping method for acquiring domain-specific features in

Sentiment Analysis.

Sentiment topic identification in English news articles

Exploiting dependency relation between sentiment topics and sentiment clue words

to induce contextual polarities of sentiment words.

Clustering-based sentiment polarity determination

Opinion Target Identification (Aug. 2008 ~ Dec. 2008)

Presented at the Multilingual Opinion Analysis Track in NTCIR-7.

Topic related opinion extraction and its target identification in English news.

KL-divergence based seed feature selection for SVM opinion extractor.

Topic language model smoothed by web-snippets to detect a topic relevance of

extracted opinion sentence to the overall topic.

Exploiting syntactic path and dependency information from opinion seeds to opinion

target for target identification.

Using Sentiment Information for Contextual Advertising (May 2008 ~ Feb. 2009)

Funded by Electronics and Telecommunications Research Institute (ETRI), Korea.

Identifying opinion target and its sentiment polarity in Korean news articles.

Filtering out ads which are relevant to negative opinions and placing ads if they are

related to positive opinions.

Language model and sentiment clue based target detection complemented by ad-

specific lexical knowledge (i.e., bid-phrase language model).

Expanding initial sentiment clues by bootstrapping utilizing collocation and the same

contextual intuition (i.e., in the same topic context, different opinion holders would

maintain the same polarities for the same target with different sentiment clues).

Exploiting a Large-Scale Commonsense Knowledge Base for Context-Driven Ad

Placement (March 2008 ~ March 2009)

Granted by Microsoft 2007 Global RFP (Request for proposal).

Funded by Microsoft Research and Microsoft adCenter.

Commonsense-aided user-context resolution and pragmatics-based approach for

contextual advertising.

Extracting ad-related commonsense knowledge from ConceptNet.

Open directory (general taxonomy) based semantic distance measurement between

the relevant category (which the extracted commonsense belong to) and the query

category (which user-query is related to).

Ranking relevant ads (shortest ads from the query by the semantic distance) by

matching keywords from the commonsense category and the query keywords (from

the query category).

Technology Trend Analysis (March 2008 August 2008)

Presented at SAC 09.

Funded by Microsoft Research Asia.

Developing semi-automatic application to detect and analyze technological trends in

US patent documents.

Curriculum Vitae Youngho Kim (abpavw@r.postjobfree.com)

Defining Technology as an association of Problem and Solution.

Problem phrase extraction using language model exploiting reference-based

document graphs and lexical patterns.

Solution phrase extraction using head words (keywords in target phrases) and

generalized linguistic patterns.

Statistical classifier-based extraction featuring language model probability, lexico-

syntactic patterns, and patent structure.

Salient technology discovery by comparing the difference between two different

language models (from the different time spans).

Paraphrasing Identification (Aug. 2007 ~ Feb. 2008)

Internship research at Microsoft Research Asia.

Syntactic structure learning for double-negation transformation.

Lexical relation (antonym and synonym) mining for light-verb construction.

Antonym acquisition for converse word substitution

Lexico-syntactic pattern and co-occurrence based bootstrapping for lexical

relationship mining

Opinion Holder Identification in English News text (March 2007 ~ July 2007)

Presented at the IEEE conference on GrC 07.

Keyword-spotting based opinionated sentence extraction and its opinion holder

identification considering anaphor problem.

Cohesive intuition (continuously holders and latent holders would appear in a

paragraph scope) based anaphoric holder resolution.

ME based multiple classifiers featuring lexical and syntactic clues.

Clue-based Opinion Analysis (Oct. 2006 ~ May 2007)

Presented at NTCIR-6 Opinion Analysis Pilot Task and AIRS 08.

Opinion mining: opinion sentence extraction, sentiment polarity determination, and

opinion holder identification in English news articles.

Lexical clues based machine learner for opinion extraction.

Rule based clue expansion and gradual gathering of training examples.

Sentiment clue and word distance based polarity determination.

Finding opinion holders using syntactic patterns.

Sentiment Analysis for Korean Economic News (Aug. 2006 ~ Jan. 2007)

Funded by Korea Telecom (KT), Korea.

Developing a domain-specific news explorer for Korean Economic news classified

up to news s sentiment polarity.

Domain-specific sentiment knowledge acquisition by labored tagging.

Comparing domain-specific lexical knowledge based approach and statistical learner

featuring general sentiment clues.

Achieving Commonsense-Based Context-awareness through Development of Korean

ConceptNet (Aug. 2006 ~ July 2007)

Presented at CICLing 07.

Funded by MIT Media Lab and Korea Science and Engineering Foundation

(KOSEF), Korea.

Semi-automatic construction of Korean specific commonsense knowledge base and

Curriculum Vitae Youngho Kim (abpavw@r.postjobfree.com)

context-aware application.

Converting large-scale English ConceptNet to Korean one by two-level translation:

sentence level and phrase level.

Visualizing Korean concept network which represents predefined semantic

relationships between concepts.

Developing a web application tool which provides a visualization of concept network

and helps users to examine the quality of automatically generated concept networks.

Designing commonsense based script for home-working robots.

Best Student Paper, Mobile-agent based home-delivery service framework In the

Awards and

Honors 1st paper contest for Logistics Innovation, sponsored by INHA University, 2005.

Bronze medal at the 2nd national collegiate programming competition, sponsored

by KAIST, Korean Institute of Information Scientists and Engineers, and Ministry

of Information and Communications, Korea, 2002.

INHA University Admission Scholarship. INHA University, 2002.

The top seat of Undergraduate Admission in Department of Computer Science and

Engineering, INHA University, 2002.

*Additional information will be provided upon request

Contact this candidate