# InterpNET: Neural Introspection for Interpretable Deep Learning

arXiv: https://arxiv.org/abs/1710.09511 · HTML: https://mlsanity.com/paper/1710.09511/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 26 Oct 2017.

## Built on by (most cited first, 5 of 5)

- [Multimodal Explainable Artificial Intelligence: A Comprehensive Review of Methodological Advances and Future Research Directions](https://mlsanity.com/paper/2306.05731/) (2023)
- [A New Perspective on Evaluation Methods for Explainable Artificial Intelligence (XAI)](https://mlsanity.com/paper/2307.14246/) (2023)
- [Explainable Authorship Identification in Cultural Heritage Applications: Analysis of a New Perspective](https://mlsanity.com/paper/2311.02237/) (2023)
- [Faithful Attention Explainer: Verbalizing Decisions Based on Discriminative Features](https://mlsanity.com/paper/2405.13032/) (2024)
- [Textual Explanations and Their Evaluations for Reinforcement Learning Policy](https://mlsanity.com/paper/2601.02514/) (2026)

## Abstract (arXiv)

Humans are able to explain their reasoning. On the contrary, deep neural networks are not. This paper attempts to bridge this gap by introducing a new way to design interpretable neural networks for classification, inspired by physiological evidence of the human visual system's inner-workings. This paper proposes a neural network design paradigm, termed InterpNET, which can be combined with any existing classification architecture to generate natural language explanations of the classifications. The success of the module relies on the assumption that the network's computation and reasoning is represented in its internal layer activations. While in principle InterpNET could be applied to any existing classification architecture, it is evaluated via an image classification and explanation task. Experiments on a CUB bird classification and explanation dataset show qualitatively and quantitatively that the model is able to generate high-quality explanations. While the current state-of-the-art METEOR score on this dataset is 29.2, InterpNET achieves a much higher METEOR score of 37.9.