Dipesh Gautam - Academia.edu (original) (raw)

Papers by Dipesh Gautam

Research paper thumbnail of Assessing Free Student Answers in Tutorial Dialogues Using LSTM Models

In this paper, we present an LSTM approach to assess free short answers in tutorial dialogue cont... more In this paper, we present an LSTM approach to assess free short answers in tutorial dialogue contexts. A major advantage of the proposed method is that it does not require any sort of feature engineering. The method performs on par and even slightly better than existing state-of-the-art methods that rely on expert-engineered features.

Research paper thumbnail of Effect of Domain Corpus Size and LSA Vector Dimension: A Study in Assessing Student Generated Short Texts in Virtual Internships Without Participant Data

Research paper thumbnail of Using Neural Tensor Networks for Open Ended Short Answer Assessment

In this paper, we present a novel approach to leverage the power of Neural Tensor Networks (NTN) ... more In this paper, we present a novel approach to leverage the power of Neural Tensor Networks (NTN) for student answer assessment in intelligent tutoring systems. The approach was evaluated on data collected using a dialogue based intelligent tutoring system (ITS). Particularly, we have experimented with different assessment models that were trained using features generated from knowledge graph embeddings derived with NTN. Our experiments showed that the model trained with the feature vectors generated with NTN, when trained with a combination of domain specific and domain general triplets, performs better than a previously proposed LSTM based approach.

Research paper thumbnail of A Bandwidth-Conserving Architecture for Crawling Virtual Worlds

A virtual world is a computer-based simulated environment intended for its users to inhabit via a... more A virtual world is a computer-based simulated environment intended for its users to inhabit via avatars. Content in virtual worlds such as Second Life or OpenSimulator is increasingly presented using three-dimensional (3D) dynamic presentation technologies that challenge traditional search technologies. As 3D environments become both more prevalent and more fragmented, the need for a data crawler and distributed search service will continue to grow. By increasing the visibility of content across virtual world servers in order to better collect and integrate the 3D data we can also improve the crawling and searching efficiency and accuracy by avoiding crawling unchanged regions or downloading unmodified objects that already exist in our collection. This will help to save bandwidth resources and Internet traffic during the content collection and indexing and, for a fixed amount of bandwidth, maximize the freshness of the collection. This work presents a new services paradigm for virtu...

Research paper thumbnail of Long Short Term Memory Based Models for Negation Handling in Tutorial Dialogues

Negation plays a significant role in spoken and written natural languages. Negation is used in la... more Negation plays a significant role in spoken and written natural languages. Negation is used in language to deny something or to reverse the polarity or the sense of a statement. This paper presents a novel approach to automatically handling negation in tutorial dialogues using deep learning methods. In particular, we explored various Long Short Term Memory (LSTM) models to automatically detect negation focus, scope and cue in tutorial dialogues collected from experiments with actual students interacting with the state-of-the-art intelligent tutoring system, DeepTutor. The results obtained are promising.

Research paper thumbnail of Combining Word Representations for Measuring Word Relatedness and Similarity

Many unsupervised methods, such as Latent Semantic Analysis and Latent Dirichlet Allocation, have... more Many unsupervised methods, such as Latent Semantic Analysis and Latent Dirichlet Allocation, have been proposed to automatically infer word representations in the form of a vector. By representing a word by a vector, one can exploit the power of vector algebra to solve many Natural Language Processing tasks e.g. by computing the cosine similarity between the corresponding word vectors the semantic similarity between the two words can be captured. In this paper, we hypothesize that combining different word representations complements the coverage of semantic aspects of a word and thus better represents the word than the individual representations. To this end, we present two approaches of combining word representations obtained from many heterogeneous sources. We also report empirical results for word-to-word semantic similarity and relatedness by using the new representation using two existing benchmark datasets.

Research paper thumbnail of Markov Analysis of Students' Professional Skills in Virtual Internships

In this paper, we conduct a Markov analysis of learners’ professional skill development based on ... more In this paper, we conduct a Markov analysis of learners’ professional skill development based on their conversations in virtual internships, an emerging category of learning systems characterized by the epistemic frame theory. This theory claims that professionals develop epistemic frames, or the network of skills, knowledge, identity, values, and epistemology (SKIVE) that are unique to that profession. Our goal here is to model individual students’ development of epistemic frames as Markov processes and infer the stationary distribution of this process, i.e. of the SKIVE elements. Our analysis of a dataset from the engineering virtual internship Nephrotex showed that domain specific SKIVE elements have higher probability. Furthermore, while comparing the SKIVE stationary distributions of pairs of individual students and display the results as heat maps, we can identify students that play leadership or coordinator roles.

Research paper thumbnail of Understanding when students are active-in-thinking through modeling-in-context

Research paper thumbnail of DT_Team at SemEval-2017 Task 1: Semantic Similarity Using Alignments, Sentence-Level Embeddings and Gaussian Mixture Model Output

Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval-2017)

We describe our system (DT Team) submitted at SemEval-2017 Task 1, Semantic Textual Similarity (S... more We describe our system (DT Team) submitted at SemEval-2017 Task 1, Semantic Textual Similarity (STS) challenge for English (Track 5). We developed three different models with various features including similarity scores calculated using word and chunk alignments, word/sentence embeddings, and Gaussian Mixture Model (GMM). The correlation between our system's output and the human judgments were up to 0.8536, which is more than 10% above baseline, and almost as good as the best performing system which was at 0.8547 correlation (the difference is just about 0.1%). Also, our system produced leading results when evaluated with a separate STS benchmark dataset. The word alignment and sentence embeddings based features were found to be very effective.

Research paper thumbnail of Evaluation Dataset (DT-Grade) and Word Weighting Approach towards Constructed Short Answers Assessment in Tutorial Dialogue Context

Proceedings of the 11th Workshop on Innovative Use of NLP for Building Educational Applications, 2016

Evaluating student answers often requires contextual information, such as previous utterances in ... more Evaluating student answers often requires contextual information, such as previous utterances in conversational tutoring systems. For example, students use coreferences and write elliptical responses, i.e. incomplete but can be interpreted in context. The DT-Grade corpus which we present in this paper consists of short constructed answers extracted from tutorial dialogues between students and an Intelligent Tutoring System and annotated for their correctness in the given context and whether the contextual information was useful. The dataset contains 900 answers (of which about 25% required contextual information to properly interpret them). We also present a baseline system developed to predict the correctness label (such as correct, correct but incomplete) in which weights for the words are assigned based on context.

Research paper thumbnail of DTSim at SemEval-2016 Task 1: Semantic Similarity Model Including Multi-Level Alignment and Vector-Based Compositional Semantics

Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), 2016

In this paper we describe our system (DT-Sim) submitted at SemEval-2016 Task 1: Semantic Textual ... more In this paper we describe our system (DT-Sim) submitted at SemEval-2016 Task 1: Semantic Textual Similarity (STS Core). We developed Support Vector Regression model with various features including the similarity scores calculated using alignment based methods and semantic composition based methods. The correlations between our system output and the human ratings were above 0.8 in three datasets.

Research paper thumbnail of Assessing Free Student Answers in Tutorial Dialogues Using LSTM Models

In this paper, we present an LSTM approach to assess free short answers in tutorial dialogue cont... more In this paper, we present an LSTM approach to assess free short answers in tutorial dialogue contexts. A major advantage of the proposed method is that it does not require any sort of feature engineering. The method performs on par and even slightly better than existing state-of-the-art methods that rely on expert-engineered features.

Research paper thumbnail of Effect of Domain Corpus Size and LSA Vector Dimension: A Study in Assessing Student Generated Short Texts in Virtual Internships Without Participant Data

Research paper thumbnail of Using Neural Tensor Networks for Open Ended Short Answer Assessment

In this paper, we present a novel approach to leverage the power of Neural Tensor Networks (NTN) ... more In this paper, we present a novel approach to leverage the power of Neural Tensor Networks (NTN) for student answer assessment in intelligent tutoring systems. The approach was evaluated on data collected using a dialogue based intelligent tutoring system (ITS). Particularly, we have experimented with different assessment models that were trained using features generated from knowledge graph embeddings derived with NTN. Our experiments showed that the model trained with the feature vectors generated with NTN, when trained with a combination of domain specific and domain general triplets, performs better than a previously proposed LSTM based approach.

Research paper thumbnail of A Bandwidth-Conserving Architecture for Crawling Virtual Worlds

A virtual world is a computer-based simulated environment intended for its users to inhabit via a... more A virtual world is a computer-based simulated environment intended for its users to inhabit via avatars. Content in virtual worlds such as Second Life or OpenSimulator is increasingly presented using three-dimensional (3D) dynamic presentation technologies that challenge traditional search technologies. As 3D environments become both more prevalent and more fragmented, the need for a data crawler and distributed search service will continue to grow. By increasing the visibility of content across virtual world servers in order to better collect and integrate the 3D data we can also improve the crawling and searching efficiency and accuracy by avoiding crawling unchanged regions or downloading unmodified objects that already exist in our collection. This will help to save bandwidth resources and Internet traffic during the content collection and indexing and, for a fixed amount of bandwidth, maximize the freshness of the collection. This work presents a new services paradigm for virtu...

Research paper thumbnail of Long Short Term Memory Based Models for Negation Handling in Tutorial Dialogues

Negation plays a significant role in spoken and written natural languages. Negation is used in la... more Negation plays a significant role in spoken and written natural languages. Negation is used in language to deny something or to reverse the polarity or the sense of a statement. This paper presents a novel approach to automatically handling negation in tutorial dialogues using deep learning methods. In particular, we explored various Long Short Term Memory (LSTM) models to automatically detect negation focus, scope and cue in tutorial dialogues collected from experiments with actual students interacting with the state-of-the-art intelligent tutoring system, DeepTutor. The results obtained are promising.

Research paper thumbnail of Combining Word Representations for Measuring Word Relatedness and Similarity

Many unsupervised methods, such as Latent Semantic Analysis and Latent Dirichlet Allocation, have... more Many unsupervised methods, such as Latent Semantic Analysis and Latent Dirichlet Allocation, have been proposed to automatically infer word representations in the form of a vector. By representing a word by a vector, one can exploit the power of vector algebra to solve many Natural Language Processing tasks e.g. by computing the cosine similarity between the corresponding word vectors the semantic similarity between the two words can be captured. In this paper, we hypothesize that combining different word representations complements the coverage of semantic aspects of a word and thus better represents the word than the individual representations. To this end, we present two approaches of combining word representations obtained from many heterogeneous sources. We also report empirical results for word-to-word semantic similarity and relatedness by using the new representation using two existing benchmark datasets.

Research paper thumbnail of Markov Analysis of Students' Professional Skills in Virtual Internships

In this paper, we conduct a Markov analysis of learners’ professional skill development based on ... more In this paper, we conduct a Markov analysis of learners’ professional skill development based on their conversations in virtual internships, an emerging category of learning systems characterized by the epistemic frame theory. This theory claims that professionals develop epistemic frames, or the network of skills, knowledge, identity, values, and epistemology (SKIVE) that are unique to that profession. Our goal here is to model individual students’ development of epistemic frames as Markov processes and infer the stationary distribution of this process, i.e. of the SKIVE elements. Our analysis of a dataset from the engineering virtual internship Nephrotex showed that domain specific SKIVE elements have higher probability. Furthermore, while comparing the SKIVE stationary distributions of pairs of individual students and display the results as heat maps, we can identify students that play leadership or coordinator roles.

Research paper thumbnail of Understanding when students are active-in-thinking through modeling-in-context

Research paper thumbnail of DT_Team at SemEval-2017 Task 1: Semantic Similarity Using Alignments, Sentence-Level Embeddings and Gaussian Mixture Model Output

Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval-2017)

We describe our system (DT Team) submitted at SemEval-2017 Task 1, Semantic Textual Similarity (S... more We describe our system (DT Team) submitted at SemEval-2017 Task 1, Semantic Textual Similarity (STS) challenge for English (Track 5). We developed three different models with various features including similarity scores calculated using word and chunk alignments, word/sentence embeddings, and Gaussian Mixture Model (GMM). The correlation between our system's output and the human judgments were up to 0.8536, which is more than 10% above baseline, and almost as good as the best performing system which was at 0.8547 correlation (the difference is just about 0.1%). Also, our system produced leading results when evaluated with a separate STS benchmark dataset. The word alignment and sentence embeddings based features were found to be very effective.

Research paper thumbnail of Evaluation Dataset (DT-Grade) and Word Weighting Approach towards Constructed Short Answers Assessment in Tutorial Dialogue Context

Proceedings of the 11th Workshop on Innovative Use of NLP for Building Educational Applications, 2016

Evaluating student answers often requires contextual information, such as previous utterances in ... more Evaluating student answers often requires contextual information, such as previous utterances in conversational tutoring systems. For example, students use coreferences and write elliptical responses, i.e. incomplete but can be interpreted in context. The DT-Grade corpus which we present in this paper consists of short constructed answers extracted from tutorial dialogues between students and an Intelligent Tutoring System and annotated for their correctness in the given context and whether the contextual information was useful. The dataset contains 900 answers (of which about 25% required contextual information to properly interpret them). We also present a baseline system developed to predict the correctness label (such as correct, correct but incomplete) in which weights for the words are assigned based on context.

Research paper thumbnail of DTSim at SemEval-2016 Task 1: Semantic Similarity Model Including Multi-Level Alignment and Vector-Based Compositional Semantics

Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), 2016

In this paper we describe our system (DT-Sim) submitted at SemEval-2016 Task 1: Semantic Textual ... more In this paper we describe our system (DT-Sim) submitted at SemEval-2016 Task 1: Semantic Textual Similarity (STS Core). We developed Support Vector Regression model with various features including the similarity scores calculated using alignment based methods and semantic composition based methods. The correlations between our system output and the human ratings were above 0.8 in three datasets.