Differentiable reimplementations of the Atari VCS provide a complex, fully known ground-truth system for testing gradient-based explainable AI methods.
StellaTeam.2024
11 Pith papers cite this work. Polarity classification is still indexing.
11
Pith papers citing it
representative citing papers
ViperGPT generates executable Python code to compose pre-trained vision-and-language modules into programs that answer visual queries, reaching state-of-the-art results with no additional training.
TIRA attacks with PMiS and PRSMP push fairness metrics to ideal values and reduce SHAP attribution for protected features to zero in black-box settings.
Adding register tokens to Vision Transformers eliminates high-norm background artifacts and raises state-of-the-art performance on dense visual prediction tasks.
VP2O maps PPO to SVGD in a MoE architecture using functional kernels and expert orthogonalization, claiming +179 ELO on Codeforces and 32% token reduction on AIME for a 33B/4B model.
OPTIMUS generates minimal and sufficient concept-based visual explanations for deep classifiers using prime implicant theory to enforce logical sufficiency and minimality.
Inpainting auxiliary task improves clustering of embeddings for individual zebrafish identification based on skin patterns.
An edge-weighted multi-graph GNN (E-PCN/KIGNet) that encodes four jet kinematic variables improves JetClass jet-tagging accuracy and attributes most predictions to angular separation and transverse momentum.
Shapley value and variational importance switch methods produce consistent rankings of filter importance in CNNs, enabling compression and interpretability.
Transfer learning with a Zoobot CNN on SDSS DR18 data identifies 3,679 lopsided spiral galaxies at 87% test accuracy, with lopsided systems showing higher star formation, bluer colors, lower mass and concentration.
citing papers explorer
-
A Differentiable Atari VCS:A Complex, Fully Known Ground Truth for Explainable AI
Differentiable reimplementations of the Atari VCS provide a complex, fully known ground-truth system for testing gradient-based explainable AI methods.
-
ViperGPT: Visual Inference via Python Execution for Reasoning
ViperGPT generates executable Python code to compose pre-trained vision-and-language modules into programs that answer visual queries, reaching state-of-the-art results with no additional training.
-
The Unseen Hand: Manipulating Model Fairness and SHAP with Targeted Identity Re-Association Attacks
TIRA attacks with PMiS and PRSMP push fairness metrics to ideal values and reduce SHAP attribution for protected features to zero in black-box settings.
-
Vision Transformers Need Registers
Adding register tokens to Vision Transformers eliminates high-norm background artifacts and raises state-of-the-art performance on dense visual prediction tasks.
-
Variational Proximal Policy Optimization
VP2O maps PPO to SVGD in a MoE architecture using functional kernels and expert orthogonalization, claiming +179 ELO on Codeforces and 32% token reduction on AIME for a 33B/4B model.
-
OPTIMUS-Prime: Minimal and Sufficient Concept Explanations for Deep Vision Models
OPTIMUS generates minimal and sufficient concept-based visual explanations for deep classifiers using prime implicant theory to enforce logical sufficiency and minimality.
-
Exploring Clustering Capability of Inpainting Model Embeddings for Pattern-based Individual Identification
Inpainting auxiliary task improves clustering of embeddings for individual zebrafish identification based on skin patterns.
-
KIGNet: Physics-Motivated Multi-Graph Representation Learning for Explainable Jet Tagging
An edge-weighted multi-graph GNN (E-PCN/KIGNet) that encodes four jet kinematic variables improves JetClass jet-tagging accuracy and attributes most predictions to angular separation and transverse momentum.
-
Neuron ranking -- an informed way to condense convolutional neural networks architecture
Shapley value and variational importance switch methods produce consistent rankings of filter importance in CNNs, enabling compression and interpretability.
-
Identifying lopsidedness in spiral galaxies using a Deep Convolutional Neural Network
Transfer learning with a Zoobot CNN on SDSS DR18 data identifies 3,679 lopsided spiral galaxies at 87% test accuracy, with lopsided systems showing higher star formation, bluer colors, lower mass and concentration.
- MalariAI: A Label-Resilient Decoupled Framework for Annotation-Agnostic Cell Segmentation and Explainable Stage Classification in Dense Malaria Blood Smears