arXiv subjects

Jihie Kim

Publications and source records attributed to Jihie Kim.

40 records · Page 3Linked to original sources

arXiv2018

A Study on Dialogue Reward Prediction for Open-Ended Conversational Agents

The amount of dialogue history to include in a conversational agent is often underestimated and/or set in an empirical and thus possibly naive way. This suggests that principled investigations into optimal context windows are urgently needed given that the amount of dialogue history and corresponding representations can play an important role in the overall performance of a conversational system. This paper studies the amount of history required by conversational agents for reliably predicting dialogue rewards. The task of dialogue reward prediction is chosen for investigating the effects of varying amounts of dialogue history and their impact on system performance. Experimental results using a dataset of 18K human-human dialogues report that lengthy dialogue histories of at least 10 sentences are preferred (25 sentences being the best in our experiments) over short ones, and that lengthy histories are useful for training dialogue reward predictors with strong positive correlations between target dialogue rewards and predicted ones.

cs.CL↗

arXiv2017

Syllable-level Neural Language Model for Agglutinative Language

Language models for agglutinative languages have always been hindered in past due to myriad of agglutinations possible to any given word through various affixes. We propose a method to diminish the problem of out-of-vocabulary words by introducing an embedding derived from syllables and morphemes which leverages the agglutinative property. Our model outperforms character-level embedding in perplexity by 16.87 with 9.50M parameters. Proposed method achieves state of the art performance over existing input prediction methods in terms of Key Stroke Saving and has been commercialized.

cs.CL↗

arXiv2017

An Embedded Deep Learning based Word Prediction

Recent developments in deep learning with application to language modeling have led to success in tasks of text processing, summarizing and machine translation. However, deploying huge language models for mobile device such as on-device keyboards poses computation as a bottle-neck due to their puny computation capacities. In this work we propose an embedded deep learning based word prediction method that optimizes run-time memory and also provides a real time prediction environment. Our model size is 7.40MB and has average prediction time of 6.47 ms. We improve over the existing methods for word prediction in terms of key stroke savings and word prediction rate.

cs.CL↗

arXiv2011

Analyzing Answers in Threaded Discussions using a Role-Based Information Network

Online discussion boards are an important medium for collaboration. The goal of our work is to understand how messages and individual discussants contribute to Q&A discussions. We present a novel network model for capturing in-formation roles of messages and discussants, and show how we identify useful answers to the initial question. We first classify information seeking or information providing roles of messages, such as question, answer or acknowledgement. We also identify user intent in the discussion as an information seeker or a provider. We capture such role information within a reply-to discussion network, and identify messages that answer seeker questions and how answeres are acknowledged. Message influences are analyzed using B-centrality measures. User influences across different threads are combined with message influences. We use the combined score in identifying the most useful answer in the thread. The resulting ranks correlate with human provided ranks with an MRR score of 0.67.

cs.SI↗