Goto

Collaborating Authors

The compelling case for descriptive analytics: sentiment analysis and natural language

#artificialintelligence

Rapid advancements in predictive and prescriptive analytics have seemingly surpassed the overall utility of descriptive analytics. But as we strive to determine what will happen, and to prepare accordingly using technologies like machine learning, it is easy to forget the main value proposition of descriptive analytics which, although less celebrated, continues to endure. Descriptive analytics doesn't reveal what might happen, what should happen, or what your plan of action should be. Instead, it illustrates something much more concrete--what actually did happen and, with the proper analysis, what to do to get the most advantageous outcome out of a situation. Sentiment analysis is perhaps one of the most pervasive use cases for descriptive analytics today.


The Definitive Guide to Natural Language Processing

#artificialintelligence

'Volkswagen's new CEO Matthias Mueller has his work cut out for him. Mueller has spent most of his career at the Volkswagen group so he knows the inner workings of the company. Now experts say he'll have to make some big, bold changes to get the largest automaker in the world back on track.' In the example above, if only the third sentence is retained by the system, the reader will certainly ask himself who is the "he" that the summary is talking about. On the other hand, an abstraction-based approach implies text generation: the summarizer does not copy text from the input but writes in its own words what it understands from the text.


Natural Language Processing for Stocks News Analysis

#artificialintelligence

In this hands-on project, we will train a Long Short Term Memory (LSTM) deep learning model to perform stocks sentiment analysis. Natural language processing (NLP) works by converting words (text) into numbers, these numbers are then used to train an AI/ML model to make predictions. In this project, we will build a machine learning model to analyze thousands of Twitter tweets to predict people's sentiment towards a particular company or stock. The algorithm could be used automatically understand the sentiment from public tweets, which could be used as a factor while making buy/sell decision of securities. Note: This course works best for learners who are based in the North America region.


SentiPrompt: Sentiment Knowledge Enhanced Prompt-Tuning for Aspect-Based Sentiment Analysis

arXiv.org Artificial Intelligence

Aspect-based sentiment analysis (ABSA) is an emerging fine-grained sentiment analysis task that aims to extract aspects, classify corresponding sentiment polarities and find opinions as the causes of sentiment. The latest research tends to solve the ABSA task in a unified way with end-to-end frameworks. Yet, these frameworks get fine-tuned from downstream tasks without any task-adaptive modification. Specifically, they do not use task-related knowledge well or explicitly model relations between aspect and opinion terms, hindering them from better performance. In this paper, we propose SentiPrompt to use sentiment knowledge enhanced prompts to tune the language model in the unified framework. We inject sentiment knowledge regarding aspects, opinions, and polarities into prompt and explicitly model term relations via constructing consistency and polarity judgment templates from the ground truth triplets. Experimental results demonstrate that our approach can outperform strong baselines on Triplet Extraction, Pair Extraction, and Aspect Term Extraction with Sentiment Classification by a notable margin.


CASA: Conversational Aspect Sentiment Analysis for Dialogue Understanding

Journal of Artificial Intelligence Research

Dialogue understanding has always been a bottleneck for many conversational tasks, such as dialogue response generation and conversational question answering. To expedite the progress in this area, we introduce the task of conversational aspect sentiment analysis (CASA) that can provide useful fine-grained sentiment information for dialogue understanding and planning. Overall, this task extends the standard aspect-based sentiment analysis to the conversational scenario with several major adaptations. To aid the training and evaluation of data-driven methods, we annotate 3,000 chit-chat dialogues (27,198 sentences) with fine-grained sentiment information, including all sentiment expressions, their polarities and the corresponding target mentions. We also annotate an out-of-domain test set of 200 dialogues for robustness evaluation. Besides, we develop multiple baselines based on either pretrained BERT or self-attention for preliminary study. Experimental results show that our BERT-based model has strong performances for both in-domain and out-of-domain datasets, and thorough analysis indicates several potential directions for further improvements.