| dc.identifier.uri | http://hdl.handle.net/11401/77242 | |
| dc.description.sponsorship | This work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree. | en_US |
| dc.format | Monograph | |
| dc.format.medium | Electronic Resource | en_US |
| dc.language.iso | en_US | |
| dc.publisher | The Graduate School, Stony Brook University: Stony Brook, NY. | |
| dc.type | Thesis | |
| dcterms.abstract | Knowledge about processes is essential for AI systems to understand and reason about the real world events. And, the systems need some form of semantic representation to perform reasoning. At the simplest level, even knowing which class of entities play key roles can be helpful in recognizing and reasoning about events. For instance, given a description ``a puddle drying in the sun", one can recognize this as an instance of the evaporation process using simple role knowledge which asserts (among other things) that the undergoer is a kind of liquid (the puddle), and the enabler is a heat source (the sun). In this work, we explore two forms of process knowledge representations, a frame representation with a fixed set of roles and a matrix representation. We developed a fully feature engineered and a non engineered (using deep LSTM) system for role classification. We improve these process knowledge extraction models by performing cross-sentence inference--over role classifier scores--which extends the standard within sentence joint inference to inference across multiple sentences. We also present our preliminary work on modeling processes as operators. | |
| dcterms.available | 2017-09-20T16:52:15Z | |
| dcterms.contributor | Schwartz, Andrew H | en_US |
| dcterms.contributor | Balasubramanian, Niranjan | en_US |
| dcterms.contributor | Yogatama, Dani. | en_US |
| dcterms.creator | Naik, Chetan | |
| dcterms.dateAccepted | 2017-09-20T16:52:15Z | |
| dcterms.dateSubmitted | 2017-09-20T16:52:15Z | |
| dcterms.description | Department of Computer Science | en_US |
| dcterms.extent | 55 pg. | en_US |
| dcterms.format | Application/PDF | en_US |
| dcterms.format | Monograph | |
| dcterms.identifier | http://hdl.handle.net/11401/77242 | |
| dcterms.issued | 2016-12-01 | |
| dcterms.language | en_US | |
| dcterms.provenance | Made available in DSpace on 2017-09-20T16:52:15Z (GMT). No. of bitstreams: 1
Naik_grad.sunysb_0771M_13011.pdf: 351058 bytes, checksum: 7eab6a357489266dadb5371d6c98538c (MD5)
Previous issue date: 1 | en |
| dcterms.publisher | The Graduate School, Stony Brook University: Stony Brook, NY. | |
| dcterms.subject | Computer science -- Artificial intelligence -- Linguistics | |
| dcterms.subject | Computational Linguistics, Deep Learning, Machine Learning, Natural Language Processing, NLP, Semantic Representation | |
| dcterms.title | An Exploratory Study on Process Representations | |
| dcterms.type | Thesis | |
