An official website of the United States government
Large language models have demonstrated impressive performance on natural language tasks, but their decision-making processes remain opaque. Existing explanation methods either suffer from limited faithfulness to the model's reasoning or produce explanations that are difficult for humans to understand. To address these challenges, we propose ProtoSurE, a novel prototype-based surrogate framework that provides faithful and understandable explanations for LLMs. ProtoSurE trains an interpretable-by-design surrogate model that aligns with the target LLM while utilizing sentence-level prototypes as understandable concepts. Extensive experiments show that ProtoSurE consistently outperforms state-of-the-art explanation methods across diverse LLMs and datasets. Importantly, ProtoSurE demonstrates strong data efficiency, requiring relatively few training examples to achieve good performance, making it practical for real-world applications.
more »
« less
Gradient Frequency Modulation for Visually Explaining Video Understanding Models
In many applications, it is essential to understand why a machine learning model makes the decisions it does, but this is inhibited by the black-box nature of state-of-the-art neural networks. Because of this, increasing attention has been paid to explainability in deep learning, including in the area of video understanding. Due to the temporal dimension of video data, the main challenge of explaining a video action recognition model is to produce spatiotemporally consistent visual explanations, which has been ignored in the existing literature. In this paper, we propose Frequency-based Extremal Perturbation (F-EP) to explain a video understanding model's decisions. Because the explanations given by perturbation methods are noisy and non-smooth both spatially and temporally, we propose to modulate the frequencies of gradient maps from the neural network model with a Discrete Cosine Transform (DCT). We show in a range of experiments that F-EP provides more spatiotemporally consistent explanations that more faithfully represent the model's decisions compared to the existing state-of-the-art methods.
more »
« less
- Award ID(s):
- 2040209
- PAR ID:
- 10358727
- Date Published:
- Journal Name:
- British Machine Vision Conference
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
While existing hierarchical text classification (HTC) methods attempt to capture label hierarchies for model training, they either make local decisions regarding each label or completely ignore the hierarchy information during inference. To solve the mismatch between training and inference as well as modeling label dependencies in a more principled way, we formulate HTC as a Markov decision process and propose to learn a Label Assignment Policy via deep reinforcement learning to determine where to place an object and when to stop the assignment process. The proposed method, HiLAP, explores the hierarchy during both training and inference time in a consistent manner and makes inter-dependent decisions. As a general framework, HiLAP can incorporate different neural encoders as base models for end-to-end training. Experiments on five public datasets and four base models show that HiLAP yields an average improvement of 33.4% in Macro-F1 over flat classifiers and outperforms state-of-the-art HTC methods by a large margin.more » « less
-
Counterfactual explanations promote explainability in machine learning models by answering the question “how should the input instance be altered to obtain a desired predicted label?". The comparison of this instance before and after perturbation can enhance human interpretation. Most existing studies on counterfactual explanations are limited in tabular data or image data. In this paper, we study the problem of counterfactual explanation generation on graphs. A few studies have explored to generate counterfactual explanations on graphs, but many challenges of this problem are still not well-addressed: 1) optimizing in the discrete and disorganized space of graphs; 2) generalizing on unseen graphs; 3) maintaining the causality in the generated counterfactuals without prior knowledge of the causal model. To tackle these challenges, we propose a novel framework CLEAR which aims to generate counterfactual explanations on graphs for graph-level prediction models. Specifically, CLEAR leverages a graph variational autoencoder based mechanism to facilitate its optimization and generalization, and promotes causality by leveraging an auxiliary variable to better identify the causal model. Extensive experiments on both synthetic and real-world graphs validate the superiority of CLEAR over state-of-the-art counterfactual explanation methods on graphs in different aspects.more » « less
-
Abstract Brain tumor is a life-threatening disease and causes about 0.25 million deaths worldwide in 2020. Magnetic Resonance Imaging (MRI) is frequently used for diagnosing brain tumors. In medically underdeveloped regions, physicians who can accurately diagnose and assess the severity of brain tumors from MRI are highly lacking. Deep learning methods have been developed to assist physicians in detecting brain tumors from MRI and determining their subtypes. In existing methods, neural architectures are manually designed by human experts, which is time-consuming and labor-intensive. To address this problem, we propose to automatically search for high-performance neural architectures for classifying brain tumors from MRIs, by leveraging a Learning-by-Self-Explanation (LeaSE) architecture search method. LeaSE consists of an explainer model and an audience model. The explainer aims at searching for a highly performant architecture by encouraging the architecture to generate high-fidelity explanations of prediction outcomes, where explanations’ fidelity is evaluated by the audience model. LeaSE is formulated as a four-level optimization problem involving a sequence of four learning stages which are conducted end-to-end. We apply LeaSE for MRI-based brain tumor classification, including four classes: glioma, meningioma, pituitary tumor, and healthy, on a dataset containing 3264 MRI images. Results show that our method can search for neural architectures that achieve better classification accuracy than manually designed deep neural networks while having fewer model parameters. For example, our method achieves a test accuracy of 90.6% and an AUC of 95.6% with 3.75M parameters while the accuracy and AUC of a human-designed network—ResNet101—is 84.5% and 90.1% respectively with 42.56M parameters. In addition, our method outperforms state-of-the-art neural architecture search methods.more » « less
-
Graph Neural Networks (GNNs) have recently been used for node and graph classification tasks with great success, but GNNs model dependencies among the attributes of nearby neighboring nodes rather than dependencies among observed node labels. In this work, we consider the task of inductive node classification using GNNs in supervised and semi-supervised settings, with the goal of incorporating label dependencies. Because current GNNs are not universal (i.e., most-expressive) graph representations, we propose a general collective learning approach to increase the representation power of any existing GNN. Our framework combines ideas from collective classification with self-supervised learning, and uses a Monte Carlo approach to sampling embeddings for inductive learning across graphs. We evaluate performance on five real-world network datasets and demonstrate consistent, significant improvement in node classification accuracy, for a variety of state-of-the-art GNNs.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Conference Paper:
- The DOI is not currently available.
