Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

HIPPD: Brain-Inspired Hierarchical Information Processing for Personality Detection

Guanming Chen, Lingzhi Shen, Xiaohao Cai, Imran Razzak, Shoaib Jameel

TL;DR

This work presents HIPPD, a brain-inspired hierarchical framework for text-based personality detection that blends a cortex-like global semantic encoder, a prefrontal-like dynamic working memory with prediction-error–driven gating, and a basal ganglia–style pool of specialised lightweight models routed by a strict winner-takes-all mechanism. Dopaminergic-like prediction-error signals regulate both memory updates and model selection, forming a closed-loop system that adapts to context and data imbalance. Across Kaggle and Pandora MBTI benchmarks, HIPPD achieves state-of-the-art Macro-F1 and accuracy, with ablations confirming the critical contributions of each component and qualitative analysis highlighting the necessity of context-driven processing over surface-word statistics. The approach offers robust performance on short, distributed signals and shows promise for extension to other psychology- and cognition-related tasks where long-range context and weak signals prevail.

Abstract

Personality detection from text aims to infer an individual's personality traits based on linguistic patterns. However, existing machine learning approaches often struggle to capture contextual information spanning multiple posts and tend to fall short in extracting representative and robust features in semantically sparse environments. This paper presents HIPPD, a brain-inspired framework for personality detection that emulates the hierarchical information processing of the human brain. HIPPD utilises a large language model to simulate the cerebral cortex, enabling global semantic reasoning and deep feature abstraction. A dynamic memory module, modelled after the prefrontal cortex, performs adaptive gating and selective retention of critical features, with all adjustments driven by dopaminergic prediction error feedback. Subsequently, a set of specialised lightweight models, emulating the basal ganglia, are dynamically routed via a strict winner-takes-all mechanism to capture the personality-related patterns they are most proficient at recognising. Extensive experiments on the Kaggle and Pandora datasets demonstrate that HIPPD consistently outperforms state-of-the-art baselines.

HIPPD: Brain-Inspired Hierarchical Information Processing for Personality Detection

TL;DR

This work presents HIPPD, a brain-inspired hierarchical framework for text-based personality detection that blends a cortex-like global semantic encoder, a prefrontal-like dynamic working memory with prediction-error–driven gating, and a basal ganglia–style pool of specialised lightweight models routed by a strict winner-takes-all mechanism. Dopaminergic-like prediction-error signals regulate both memory updates and model selection, forming a closed-loop system that adapts to context and data imbalance. Across Kaggle and Pandora MBTI benchmarks, HIPPD achieves state-of-the-art Macro-F1 and accuracy, with ablations confirming the critical contributions of each component and qualitative analysis highlighting the necessity of context-driven processing over surface-word statistics. The approach offers robust performance on short, distributed signals and shows promise for extension to other psychology- and cognition-related tasks where long-range context and weak signals prevail.

Abstract

Personality detection from text aims to infer an individual's personality traits based on linguistic patterns. However, existing machine learning approaches often struggle to capture contextual information spanning multiple posts and tend to fall short in extracting representative and robust features in semantically sparse environments. This paper presents HIPPD, a brain-inspired framework for personality detection that emulates the hierarchical information processing of the human brain. HIPPD utilises a large language model to simulate the cerebral cortex, enabling global semantic reasoning and deep feature abstraction. A dynamic memory module, modelled after the prefrontal cortex, performs adaptive gating and selective retention of critical features, with all adjustments driven by dopaminergic prediction error feedback. Subsequently, a set of specialised lightweight models, emulating the basal ganglia, are dynamically routed via a strict winner-takes-all mechanism to capture the personality-related patterns they are most proficient at recognising. Extensive experiments on the Kaggle and Pandora datasets demonstrate that HIPPD consistently outperforms state-of-the-art baselines.

Paper Structure

This paper contains 13 sections, 3 equations, 3 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Our Novel HIPPD Model
  4. Global Semantic Encoder
  5. Dynamic Working Memory Module
  6. Specialist Model Routing
  7. Personality Classification
  8. Experiments and Discussions
  9. Experimental Setup
  10. Overall Results
  11. Ablation Study
  12. Qualitative Analysis
  13. Conclusion

Figures (3)

  • Figure 1: The visualization shows the co-occurrence matrix of the top 10 most frequent words in the real texts from the Kaggle (left) and Pandora (right) datasets.
  • Figure 2: Overview of the HIPPD architecture. The model simulates three core brain modules: the cerebral cortex, the prefrontal cortex, and the basal ganglia. Dopaminergic modulation adaptively regulates working memory and specialist selection.
  • Figure 3: The visualization shows the mutual information between the top 10 most frequent words and all MBTI types in the Kaggle (left) and Pandora (right) datasets.