Goto

Collaborating Authors

 Discourse & Dialogue


Learning Topic Models: Identifiability and Finite-Sample Analysis

arXiv.org Machine Learning

Topic models provide a useful text-mining tool for learning, extracting and discovering latent structures in large text corpora. Although a plethora of methods have been proposed for topic modeling, a formal theoretical investigation on the statistical identifiability and accuracy of latent topic estimation is lacking in the literature. In this paper, we propose a maximum likelihood estimator (MLE) of latent topics based on a specific integrated likelihood, which is naturally connected to the concept of volume minimization in computational geometry. Theoretically, we introduce a new set of geometric conditions for topic model identifiability, which are weaker than conventional separability conditions relying on the existence of anchor words or pure topic documents. We conduct finite-sample error analysis for the proposed estimator and discuss the connection of our results with existing ones. We conclude with empirical studies on both simulated and real datasets.


Situated Dialogue Learning through Procedural Environment Generation

arXiv.org Artificial Intelligence

We teach goal-driven agents to interactively act and speak in situated environments by training on generated curriculums. Our agents operate in LIGHT (Urbanek et al., 2019)--a large-scale crowd-sourced fantasy text adventure game wherein an agent perceives and interacts with the world through textual natural language. Goals in this environment take the form of character-based quests, consisting of personas and motivations. We augment LIGHT by learning to procedurally generate additional novel textual worlds and quests to create a curriculum of steadily increasing difficulty for training agents to achieve such goals. In particular, we measure curriculum difficulty in terms of the rarity of the quest in the original training distribution--an easier environment is one that is more likely to have been found in the unaugmented dataset. An ablation study shows that this method of learning from the tail of a distribution results in significantly higher generalization abilities as measured by zeroshot performance on never-before-seen quests. Figure 1: The LIGHT questing environment presented as a 2 player game deployed in Messenger.


Exploring Conditional Text Generation for Aspect-Based Sentiment Analysis

arXiv.org Artificial Intelligence

Aspect-based sentiment analysis (ABSA) is an NLP task that entails processing user-generated reviews to determine (i) the target being evaluated, (ii) the aspect category to which it belongs, and (iii) the sentiment expressed towards the target and aspect pair. In this article, we propose transforming ABSA into an abstract summary-like conditional text generation task that uses targets, aspects, and polarities to generate auxiliary statements. To demonstrate the efficacy of our task formulation and a proposed system, we fine-tune a pre-trained model for conditional text generation tasks to get new state-of-the-art results on a few restaurant domains and urban neighborhoods domain benchmark datasets.


Twitter Sentiment Analysis with Python

#artificialintelligence

Since the feud between James and Tati took place in 2019, we will scrape Tweets from that time. We can do this with the help of a library called Twint. First, install this library with a simple pip intall twint . Now, let's run the following lines of code: The above lines of code will scrape 50K Tweets with the hashtag #jamescharles from January 2019. Let's now take a look at some of the variables present in the data frame: The data frame has 35 columns, and I've only attached a screenshot of half of them.


Beyond Topics: Discovering Latent Healthcare Objectives from Event Sequences

arXiv.org Artificial Intelligence

A meaningful understanding of clinical protocols and patient pathways helps improve healthcare outcomes. Electronic health records (EHR) reflect real-world treatment behaviours that are used to enhance healthcare management but present challenges; protocols and pathways are often loosely defined and with elements frequently not recorded in EHRs, complicating the enhancement. To solve this challenge, healthcare objectives associated with healthcare management activities can be indirectly observed in EHRs as latent topics. Topic models, such as Latent Dirichlet Allocation (LDA), are used to identify latent patterns in EHR data. However, they do not examine the ordered nature of EHR sequences, nor do they appraise individual events in isolation. Our novel approach, the Categorical Sequence Encoder (CaSE) addresses these shortcomings. The sequential nature of EHRs is captured by CaSE's event-level representations, revealing latent healthcare objectives. In synthetic EHR sequences, CaSE outperforms LDA by up to 37% at identifying healthcare objectives. In the real-world MIMIC-III dataset, CaSE identifies meaningful representations that could critically enhance protocol and pathway development.


A Comparative Study of Sentiment Analysis Using NLP and Different Machine Learning Techniques on US Airline Twitter Data

arXiv.org Artificial Intelligence

Today's business ecosystem has become very competitive. Customer satisfaction has become a major focus for business growth. Business organizations are spending a lot of money and human resources on various strategies to understand and fulfill their customer's needs. But, because of defective manual analysis on multifarious needs of customers, many organizations are failing to achieve customer satisfaction. As a result, they are losing customer's loyalty and spending extra money on marketing. We can solve the problems by implementing Sentiment Analysis. It is a combined technique of Natural Language Processing (NLP) and Machine Learning (ML). Sentiment Analysis is broadly used to extract insights from wider public opinion behind certain topics, products, and services. We can do it from any online available data. In this paper, we have introduced two NLP techniques (Bag-of-Words and TF-IDF) and various ML classification algorithms (Support Vector Machine, Logistic Regression, Multinomial Naive Bayes, Random Forest) to find an effective approach for Sentiment Analysis on a large, imbalanced, and multi-classed dataset. Our best approaches provide 77% accuracy using Support Vector Machine and Logistic Regression with Bag-of-Words technique.


UserIdentifier: Implicit User Representations for Simple and Effective Personalized Sentiment Analysis

arXiv.org Artificial Intelligence

Global models are trained to be as generalizable as possible, with user invariance considered desirable since the models are shared across multitudes of users. As such, these models are often unable to produce personalized responses for individual users, based on their data. Contrary to widely-used personalization techniques based on few-shot learning, we propose UserIdentifier, a novel scheme for training a single shared model for all users. Our approach produces personalized responses by adding fixed, non-trainable user identifiers to the input data. We empirically demonstrate that this proposed method outperforms the prefix-tuning based state-of-the-art approach by up to 13%, on a suite of sentiment analysis datasets. We also show that, unlike prior work, this method needs neither any additional model parameters nor any extra rounds of few-shot fine-tuning.


Sentiment Analysis in Twitter for Macedonian

arXiv.org Artificial Intelligence

We present work on sentiment analysis in Twitter for Macedonian. As this is pioneering work for this combination of language and genre, we created suitable resources for training and evaluating a system for sentiment analysis of Macedonian tweets. In particular, we developed a corpus of tweets annotated with tweet-level sentiment polarity (positive, negative, and neutral), as well as with phrase-level sentiment, which we made freely available for research purposes. We further bootstrapped several large-scale sentiment lexicons for Macedonian, motivated by previous work for English. The impact of several different pre-processing steps as well as of various features is shown in experiments that represent the first attempt to build a system for sentiment analysis in Twitter for the morphologically rich Macedonian language. Overall, our experimental results show an F1-score of 92.16, which is very strong and is on par with the best results for English, which were achieved in recent SemEval competitions.


Every time I fire a conversational designer, the performance of the dialog system goes down

arXiv.org Artificial Intelligence

Incorporating explicit domain knowledge into neural-based task-oriented dialogue systems is an effective way to reduce the need of large sets of annotated dialogues. In this paper, we investigate how the use of explicit domain knowledge of conversational designers affects the performance of neural-based dialogue systems. To support this investigation, we propose the Conversational-Logic-Injection-in-Neural-Network system (CLINN) where explicit knowledge is coded in semi-logical rules. By using CLINN, we evaluated semi-logical rules produced by a team of differently skilled conversational designers. We experimented with the Restaurant topic of the MultiWOZ dataset. Results show that external knowledge is extremely important for reducing the need of annotated examples for conversational systems. In fact, rules from conversational designers used in CLINN significantly outperform a state-of-the-art neural-based dialogue system.


ExCode-Mixed: Explainable Approaches towards Sentiment Analysis on Code-Mixed Data using BERT models

arXiv.org Artificial Intelligence

The increasing use of social media sites in countries like India has given rise to large volumes of code-mixed data. Sentiment analysis of this data can provide integral insights into people's perspectives and opinions. Developing robust explainability techniques which explain why models make their predictions becomes essential. In this paper, we propose an adequate methodology to integrate explainable approaches into code-mixed sentiment analysis.