TraceNet: Tracing and Locating the Key Elements in Sentiment Analysis
Qinghua Zhao, Shuai Ma

TL;DR
TraceNet is a neural architecture designed for sentiment analysis that identifies and traces key input elements, improving interpretability and robustness by focusing on important features and masking less relevant parts.
Contribution
The paper introduces TraceNet, a novel neural model with locators and encoders organized layer-wise, employing regularization and sparsity constraints for better interpretability and robustness in sentiment analysis.
Findings
Effective in sentiment classification tasks
Enhances interpretability by tracing key input elements
Improves robustness through proactive masking
Abstract
In this paper, we study sentiment analysis task where the outcomes are mainly contributed by a few key elements of the inputs. Motivated by the two-streams hypothesis, we propose a neural architecture, named TraceNet, to address this type of task. It not only learns discriminative representations for the target task via its encoders, but also traces key elements at the same time via its locators. In TraceNet, both encoders and locators are organized in a layer-wise manner, and a smoothness regularization is employed between adjacent encoder-locator combinations. Moreover, a sparsity constraints are enforced on locators for tracing purposes and items are proactively masked according to the item weights output by locators.A major advantage of TraceNet is that the outcomes are easier to understand, since the most responsible parts of inputs are identified. Also, under the guidance of…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsSentiment Analysis and Opinion Mining · Topic Modeling · Text and Document Classification Technologies
