Search Results for author:
Found 13 papers, 12 papers with code
B-cosification: Transforming Deep Neural Networks to be Inherently Interpretable
We find that B-cosification can yield models that are on par with B-cos models trained from scratch in terms of interpretability, while often outperforming them in terms of classification performance at a fraction of the training cost.
Discover-then-Name: Task-Agnostic Concept Bottlenecks via Automated Concept Discovery
Concept Bottleneck Models (CBMs) have recently been proposed to address the 'black-box' problem of deep neural networks, by first mapping images to a human-understandable concept space and then linearly combining concepts for classification.
Good Teachers Explain: Explanation-Enhanced Knowledge Distillation
Knowledge Distillation (KD) has proven effective for compressing large teacher models into smaller student models.
B-cos Alignment for Inherently Interpretable CNNs and Vision Transformers
We present a new direction for increasing the interpretability of deep neural networks (DNNs) by promoting weight-input alignment during training.
Temperature Schedules for Self-Supervised Contrastive Methods on Long-Tail Data
Such a schedule results in a constant `task switching' between an emphasis on instance discrimination and group-wise discrimination and thereby ensures that the model learns both group-wise features, as well as instance-specific details.
Better Understanding Differences in Attribution Methods via Systematic Evaluations
Finally, we propose a post-processing smoothing step that significantly improves the performance of some attribution methods, and discuss its applicability.
Studying How to Efficiently and Effectively Guide Models with Explanations
To better understand the effectiveness of the various design choices that have been explored in the context of model guidance, in this work we conduct an in-depth evaluation across various loss functions, attribution methods, models, and 'guidance depths' on the PASCAL VOC 2007 and MS COCO 2014 datasets.
Holistically Explainable Vision Transformers
Transformers increasingly dominate the machine learning landscape across many tasks and domains, which increases the importance for understanding their outputs.
B-cos Networks: Alignment is All We Need for Interpretability
We present a new direction for increasing the interpretability of deep neural networks (DNNs) by promoting weight-input alignment during training.
Towards Better Understanding Attribution Methods
Finally, we propose a post-processing smoothing step that significantly improves the performance of some attribution methods, and discuss its applicability.
Optimising for Interpretability: Convolutional Dynamic Alignment Networks
As a result, CoDA Nets model the classification prediction through a series of input-dependent linear transformations, allowing for linear decomposition of the output into individual input contributions.
Convolutional Dynamic Alignment Networks for Interpretable Classifications
Given the alignment of the DAUs, the resulting contribution maps align with discriminative input patterns.
Visualizing evidence for Alzheimer's disease in deep neural networks trained on structural MRI data
In this study, we propose using layer-wise relevance propagation (LRP) to visualize convolutional neural network decisions for AD based on MRI data.
2D Human Pose Estimation
Quantitative Methods
