Multi-Task Learning with LLMs for Implicit Sentiment Analysis: Data-level and Task-level Automatic Weight Learning

TL;DR

This work tackles implicit sentiment analysis (ISA) by integrating large language models (LLMs) into a multi-task learning (MTL) framework named MT-ISA. MT-ISA uses LLMs to generate auxiliary sentiment elements (aspect $a$ and opinion $o$) and applies data-level automatic weight learning (D-AWL) across three strategies (Input, Output, Input-Output) together with task-level automatic weight learning (T-AWL) based on homoscedastic uncertainty, yielding an automatic loss function (ALF) that balances auxiliary and primary tasks. Experiments on SemEval-2014 Restaurant and Laptop datasets with Flan-T5 backbones (Base to XXL) show MT-ISA achieves state-of-the-art results, and ablation studies confirm that both D-AWL and T-AWL contribute significantly to performance, with model size influencing the optimal AWL strategy. The findings demonstrate robust, scalable reasoning for ISA across model sizes, reducing manual tuning and highlighting the practical impact of uncertainty-aware MTL with LLM-generated auxiliary data.

Abstract

Implicit sentiment analysis (ISA) presents significant challenges due to the absence of salient cue words. Previous methods have struggled with insufficient data and limited reasoning capabilities to infer underlying opinions. Integrating multi-task learning (MTL) with large language models (LLMs) offers the potential to enable models of varying sizes to reliably perceive and recognize genuine opinions in ISA. However, existing MTL approaches are constrained by two sources of uncertainty: data-level uncertainty, arising from hallucination problems in LLM-generated contextual information, and task-level uncertainty, stemming from the varying capacities of models to process contextual information. To handle these uncertainties, we introduce MT-ISA, a novel MTL framework that enhances ISA by leveraging the generation and reasoning capabilities of LLMs through automatic MTL. Specifically, MT-ISA constructs auxiliary tasks using generative LLMs to supplement sentiment elements and incorporates automatic MTL to fully exploit auxiliary data. We introduce data-level and task-level automatic weight learning (AWL), which dynamically identifies relationships and prioritizes more reliable data and critical tasks, enabling models of varying sizes to adaptively learn fine-grained weights based on their reasoning capabilities. We investigate three strategies for data-level AWL, while also introducing homoscedastic uncertainty for task-level AWL. Extensive experiments reveal that models of varying sizes achieve an optimal balance between primary prediction and auxiliary tasks in MT-ISA. This underscores the effectiveness and adaptability of our approach.

Figures (6)

• Figure 1: The examples illustrate explicit (left) and implicit (right) cases in ABSA. LLMs help supplement sentiment elements in ISA.
• Figure 2: The overview of proposed multi-task learning framework MT-ISA. The primary task is polarity inference originating from the given dataset. The auxiliary tasks are constructed by LLM generation using the self-refine strategy with polarity intervention to guide the generation for relevant sentiment elements, including aspect and opinion. The backbone model is trained using multi-task learning with automatic weight learning (AWL), which simultaneously considers auxiliary data confidence for data-level AWL and homoscedastic uncertainty (i.e., task-level uncertainty) for task-level AWL to obtain fine-grained weights and achieve optimal learning performance.
• Figure 3: Comparision of performance with different D-AWL strategies, including input (I), output (O), and input-output (I-O) strategies. The metric uses the accuracy of implicit datasets.
• Figure 4: The model size effect for MT-ISA with input and output D-AWL strategies. Both auxiliary task weights and implicit F1 score are compared.
• Figure 5: An example illustrates Algorithm \ref{['refine']} designed for constructing auxiliary data. This process generates both aspect $a$ and opinion $o$ using a self-refine strategy with gold label intervention. This example reaches a consensus with GPT-4o-mini after two runs.