Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Aug 30, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

675

Multi-task learning for toxic comment classification and rationale extraction.

Kiran Babu Nelatoori1, Hima Bindu Kommanti1

  • 1Department of CSE, National Institute of Technology Andhra Pradesh, 534101 Andhra Pradesh, India.

Journal of Intelligent Information Systems
|August 29, 2022
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Toxic Reactions: Overview01:26

Toxic Reactions: Overview

1.1K
When toxic substances penetrate the human body, they disseminate to various tissues, undergoing metabolic changes. This process yields reactive metabolites that may covalently bind with specific target molecules, resulting in toxicity.
Toxicity falls into two primary categories: local and systemic.
Local toxicity appears at the exposure site, such as protein denaturation caused by caustic substances.
In contrast, systemic toxicity requires the toxic agent's absorption and distribution,...
1.1K
Enhanced Elimination of Poison01:26

Enhanced Elimination of Poison

569
Poison can be effectively removed from the gastrointestinal (GI) tract through various decontamination procedures.
Antidotes serve a crucial role in counteracting the effects of poison by inhibiting enzymes responsible for producing harmful drug metabolites. In some cases, these toxic metabolites can be neutralized by endogenous cosubstrates, which are maintained at specific concentrations to prevent interaction with cellular macromolecules and subsequent cell death.
Renal excretion is the...
569
Types of Toxins01:36

Types of Toxins

1.9K
Humans continually engage with an environment rich in potentially harmful chemicals. These are introduced to our bodies through inhalation, ingestion, or skin contact. These chemicals exist in various forms, such as air and environmental pollutants, agricultural chemicals, organic solvents, and heavy metals.
Air pollutants, primarily gases, pose significant threats to respiratory health, leading to conditions like hypoxia, lung cancer, and in extreme cases, death.
Environmental pollutants like...
1.9K
Anticholinesterase Agents: Poisoning and Treatment01:26

Anticholinesterase Agents: Poisoning and Treatment

979
Anticholinesterases, also known as cholinesterase inhibitors, work by blocking the breakdown of acetylcholine, leading to its accumulation in the synaptic cleft. This accumulation indirectly enhances both muscarinic and nicotinic actions. These agents are classified as reversible or irreversible based on their mechanism of action.     
Irreversible agents form a strong bond with the cholinesterase enzyme, making it inactive. The breakdown of the phosphorylated enzyme is...
979
Force Classification01:22

Force Classification

1.4K
Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
1.4K
Predicting Reaction Outcomes02:24

Predicting Reaction Outcomes

8.6K
Kinetics describes the rate and path by which a reaction occurs. In contrast, thermodynamics deals with state functions and describes the properties, behavior, and components of a system. It is not concerned with the path taken by the process and cannot address the rate at which a reaction occurs. Although it does provide information about what can happen during a reaction process, it does not describe the detailed steps of what appears on an atomic or a molecular level. On the other hand,...
8.6K

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same journal

Using ontologies to facilitate healthcare process mining and analysis.

Journal of intelligent information systems·2026
Same journal

Benchmarking state of the art website embedding methods for effective processing and analysis in the public sector.

Journal of intelligent information systems·2026
Same journal

Identifying multimodal misinformation leveraging novelty detection and emotion recognition.

Journal of intelligent information systems·2023
Same journal

Multilingual deep learning framework for fake news detection using capsule neural network.

Journal of intelligent information systems·2023
Same journal

Generalized durative event detection on social media.

Journal of intelligent information systems·2023
Same journal

Towards a soft three-level voting model (Soft T-LVM) for fake news detection.

Journal of intelligent information systems·2022
See all related articles

This study introduces a multi-task learning model for improved toxic comment classification and rationale identification on social media. The model enhances automated content moderation systems by jointly learning these tasks.

Area of Science:

  • Natural Language Processing
  • Computational Social Science

Background:

  • Social media content moderation is crucial for healthy online discussions.
  • Toxic span prediction aids in understanding and explaining toxic comment classifications.
  • Automated moderation systems benefit from accurate toxic span prediction.

Purpose of the Study:

  • To develop a multi-task learning (MTL) model for joint toxic comment classification and toxic span prediction.
  • To improve the performance of automated content moderation systems.
  • To evaluate the model's domain adaptation capabilities on out-of-domain datasets.

Main Methods:

  • Proposed a multi-task learning model utilizing ToxicXLMR for text embeddings and a Bi-LSTM CRF layer for rationale identification.
  • Curated a dataset combining Jigsaw and toxic span prediction datasets for multi-task learning.
Keywords:
Joint lossMulti-Task Learning (MTL)Rationale extractionToxic Comment Classification (TCC)Toxic Span Prediction (TSP)Transfer learning

More Related Videos

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
05:47

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

Published on: June 13, 2025

547
Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers
03:37

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Published on: March 1, 2024

859

Related Experiment Videos

Last Updated: Aug 30, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

675
Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
05:47

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

Published on: June 13, 2025

547
Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers
03:37

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Published on: March 1, 2024

859
  • Evaluated the model on curated, toxic span prediction, HASOC, and OLID datasets.

    • Main Results:

    • The proposed MTL model achieved 4% and 2% improvement over single-task models for classification and rationale identification, respectively.
    • Demonstrated a 3% F1 score improvement on out-of-domain datasets (HASOC, OLID) compared to single-task models.
    • The joint learning objective proved meaningful for toxic comment analysis.

    Conclusions:

    • The developed multi-task learning model effectively enhances both toxic comment classification and toxic span prediction.
    • The model shows promising domain adaptation capabilities for analyzing diverse social media text.
    • This approach contributes to building more robust automated content moderation systems.