arXiv subjects

Thomas Hoang

Publications and source records attributed to Thomas Hoang.

5 recordsLinked to original sources

arXiv2026

DeLIVeR: Decomposed Learning for Information-grounded Veracity Recognition via Reinforced Knowledge Graph Exploration

Automated fact-checking remains a challenge for Large Language Models (LLMs) due to "query brittleness" in traditional retrieval systems. We propose DeLIVeR (Decomposed Learning for Information-grounded Veracity Recognition), a framework that treats evidence retrieval as a reinforced strategic exploration task. DeLIVeR utilizes a Planner LLM to decompose complex claims into targeted question sets, which are used to traverse structured Knowledge Graphs (KGs) for high-precision evidence. We optimize the Planner's policy using Group Relative Policy Optimization (GRPO) with a reward system prioritizing structural diversity and verdict accuracy. Our evaluation on LIAR, FEVER, and PolitiFact shows that DeLIVeR significantly outperforms state-of-the-art baselines. Using Qwen2.5-7B, our framework achieved peak F1-scores of 83.73, 84.57, and 79.70 respectively, representing a 10-15% improvement over HippoRAG2. By shifting to a reinforced question-planning strategy, DeLIVeR effectively bridges multi-hop reasoning gaps and provides an auditable, transparent path for verifiable misinformation detection.

cs.CL↗

arXiv2025

Algorithms for estimating linear function in data mining

The main goal of this topic is to showcase several studied algorithms for estimating the linear utility function to predict the users preferences. For example, if a user comes to buy a car that has several attributes including speed, color, age, etc in a linear function, the algorithms that we present in this paper help with estimating this linear function to filter out a small subset that would be of best interest to the user among a million tuples in a very large database. In addition, the estimating linear function could also be applicable in getting to know what the data can do or predicting the future based on the data that is used in data science, which is demonstrated by the GNN, PLOD algorithms. In the ever-evolving field of data science, deriving valuable insights from large datasets is critical for informed decision-making, particularly in predictive applications. Data analysts often identify high-quality datasets without missing values, duplicates, or inconsistencies before merging diverse attributes for analysis. Taking housing price prediction as a case study, various attributes must be considered, including location factors (proximity to urban centers, crime rates), property features (size, style, modernity), and regional policies (tax implications). Experts in the field typically rank these attributes to establish a predictive utility function, which machine learning models use to forecast outcomes like housing prices. Several data discovery algorithms, including those that address the challenges of predefined utility functions and human input for attribute ranking, which often result in a time-consuming iterative process, that the work of cannot overcome.

cs.IR↗

arXiv2024

GNN: Graph Neural Network and Large Language Model for Data Discovery

Our algorithm GNN: Graph Neural Network and Large Language Model for Data Discovery inherit the benefits of \cite{hoang2024plod} (PLOD: Predictive Learning Optimal Data Discovery), \cite{Hoang2024BODBO} (BOD: Blindly Optimal Data Discovery) in terms of overcoming the challenges of having to predefine utility function and the human input for attribute ranking, which helps prevent the time-consuming loop process. In addition to these previous works, our algorithm GNN leverages the advantages of graph neural networks and large language models to understand text type values that cannot be understood by PLOD and MOD, thus making the task of predicting outcomes more reliable. GNN could be seen as an extension of PLOD in terms of understanding the text type value and the user's preferences, not only numerical values but also text values, making the promise of data science and analytics purposes.

cs.DB↗

arXiv2024

Public Health in Disaster: Emotional Health and Life Incidents Extraction during Hurricane Harvey

Countless disasters have resulted from climate change, causing severe damage to infrastructure and the economy. These disasters have significant societal impacts, necessitating mental health services for the millions affected. To prepare for and respond effectively to such events, it is important to understand people's emotions and the life incidents they experience before and after a disaster strikes. In this case study, we collected a dataset of approximately 400,000 public tweets related to the storm. Using a BERT-based model, we predicted the emotions associated with each tweet. To efficiently identify these topics, we utilized the Latent Dirichlet Allocation (LDA) technique for topic modeling, which allowed us to bypass manual content analysis and extract meaningful patterns from the data. However, rather than stopping at topic identification like previous methods \cite{math11244910}, we further refined our analysis by integrating Graph Neural Networks (GNN) and Large Language Models (LLM). The GNN was employed to generate embeddings and construct a similarity graph of the tweets, which was then used to optimize clustering. Subsequently, we used an LLM to automatically generate descriptive names for each event cluster, offering critical insights for disaster preparedness and response strategies.

cs.IR↗

arXiv2024

BOD: Blindly Optimal Data Discovery

Combining discovery and augmentation is important in the era of data usage when it comes to predicting the outcome of tasks. However, having to ask the user the utility function to discover the goal to achieve the optimal small rightful dataset is not an optimal solution. The existing solutions do not make good use of this combination, hence underutilizing the data. In this paper, we introduce a novel goal-oriented framework, called BOD: Blindly Optimal Data Discovery, that involves humans in the loop and comparing utility scores every time querying in the process without knowing the utility function. This establishes the promise of using BOD: Blindly Optimal Data Discovery for modern data science solutions.

cs.DB↗