Integrated Gradients

Also known as IG

Integrated Gradients attributes a prediction to input features by accumulating gradients between a chosen baseline and the input, aiming to account for the prediction difference. Work extends this approach with paths aligned to data manifolds, spectral or predictive-distribution paths, and layer-wise or time-resolved attribution for Transformers and diffusion models. Applications include image and audio explanation, histopathology, multimodal models, and causal feature analysis. A central challenge is that explanations can change with the baseline, path, model architecture, and evaluation metric, so reliable comparisons and validation remain important.

Latest papers

14 papers3 papers in the last four weeks, with none the four weeks before. 0.0% of all new papers.

Causal Integrated Gradients: Scalable Feature Attribution Along Causal Paths
Oct 15, 2026Joran Michiels, Anis Ismail, Maarten De Vos +1Causal Attribution Integrated Gradients
Laboratory of Multi-omics Integrative Bioinformatics, Department of Human Genetics, KU Leuven, Leuven, Belgium · KU Leuven Institute for Single Cell Omics (LISCO), KU Leuven, Leuven, Belgium · Department of Electrical Engineering, KU Leuven, Leuven, Belgium +1
Temporally-Resolved Token Attribution Reveals the Generation Dynamics of Diffusion Language Models
Oct 1, 2026Darpan Aswal, Céline HudelotIntegrated Gradients LLM Interpretability
Université Grenoble Alpes · MICS, CentraleSupélec
Restoring without Forgetting: Filter-Level Continual Image Restoration via Parameter-Space Integrated Gradients
Sep 29, 2026Xin Feng, Jin Zhao, Yizhen Zhang +3Integrated Gradients Image Restoration
School of Informatics, University of Edinburgh · Baidu Inc. · Tsinghua University +1
Beyond Attention: Signed Integrated Gradients Attribution in a BiomeGPT-Style Microbiome Transformer
Aug 6, 2026Oren NelsonGradient-Based Attribution Integrated Gradients
Department of Computer Science University of California, San Diego San Diego, CA 92122
IG-Lens: Exact Additive Probability Attribution Across Transformer Layers via Telescoping Integrated Gradients
Jun 29, 2026Duc Anh NguyenGradient-Based Attribution Integrated Gradients
Hanoi University of Science and Technology
Evaluating the Temporal Detection Capability of Integrated Gradients Applied on Sound Classifier
May 22, 2026Martynas Dumpis, Tuomas VirtanenGradient-Based Attribution Integrated Gradients
Department of Electronic Systems · Vilnius Gediminas Technical University · Vilnius, Lithuania +3
The Neglected Baseline in Model Interpretation
May 21, 2026Yongjin Cui, Xiaohui FanGradient-Based Attribution Integrated Gradients
Zhejiang University
Spectral Integrated Gradients for Coarse-to-Fine Feature Attribution
May 19, 2026Soyeon Kim, Seongwoo Lim, Kyowoon Lee +1Gradient-Based Attribution Integrated Gradients
Korea Advanced Institute of Science and Technology Seongnam, Gyeonggi-do, Republic of Korea · INEEJI Corp. Seongnam, Gyeonggi-do, Republic of Korea
FRInGe: Distribution-Space Integrated Gradients with Fisher--Rao Geometry
May 7, 2026Gabriele Martino, Sebastian TschiatschekGradient-Based Attribution Integrated Gradients
Faculty of Computer Science / Doctoral School Computer Science University of Vienna Vienna, Austria · Faculty of Computer Science / ds:UniVie University of Vienna Vienna, Austria
Manifold-Aligned Guided Integrated Gradients for Reliable Feature Attribution
May 4, 2026Soyeon Kim, Seongwoo Lim, Kyowoon Lee +1Gradient-Based Attribution Integrated Gradients
Kim Jaechul Graduate School of AI, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea · INEEJI, Seongnam, Korea · KAIST InnoCORE LLM, KAIST, Daejeon, Korea
Path-Sampled Integrated Gradients
Apr 15, 2026Firuz Kamalov, Fadi Thabtah, R. Sivaraj +1Gradient-Based Attribution Integrated Gradients
Department of Computational Sciences, Canadian University Dubai, UAE · Abu Dhabi School of Management, Abu Dhabi, UAE · Department of Mathematics, Dr. B.R. Ambedkar National Institute of Technology, India
Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients
Mar 18, 2026Ziwei Xiang, Fanhu Zeng, Hongjian Fang +6VLM Quantization Post-Training Quantization
State Key Laboratory of Multimodal Artificial Intelligence Systems, CASIA · School of Artificial Intelligence, UCAS · Beijing National Research Center for Information Science and Technology +3
Beyond Attention Heatmaps: How to Get Better Explanations for Multiple Instance Learning Models in Histopathology
Mar 9, 2026Mina Jamshidi Idaji, Julius Hense, Tom Neuhäuser +12Saliency Map Evaluation Multiple Instance Learning
Berlin Institute for the Foundations of Learning and Data, Berlin, Germany · Machine Learning Group, Technische Universität Berlin, Berlin, Germany · Department of Chemistry, Chemical Physics Theory Group, University of Toronto, Canada +9
Explainable AI: Context-Aware Layer-Wise Integrated Gradients for Explaining Transformer Models
Feb 18, 2026Melkamu Abay Mersha, Jugal KalitaTransformer Interpretability Gradient-Based Attribution
College of Engineering and Applied Science, University of Colorado Colorado Springs, Colorado Springs, 80918, CO, USA

Oct 15, 2026J/K move · Enter open · S save

Causal Integrated Gradients: Scalable Feature Attribution Along Causal Paths

Joran Michiels, Anis Ismail, Maarten De Vos, Alejandro Sifrim

Laboratory of Multi-omics Integrative Bioinformatics, Department of Human Genetics, KU Leuven, Leuven, Belgium · KU Leuven Institute for Single Cell Omics (LISCO), KU Leuven, Leuven, Belgium · Department of Electrical Engineering, KU Leuven, Leuven, Belgium+1

Causal Attribution Integrated Gradients

Open paper AskPDF arXiv

Existing feature attribution methods that account for feature dependencies do not scale to high dimensions. We propose causal integrated gradients (CIG) and causal expected gradients (CEG), which extend integrated gradients to incorporate prior causal structure among input features while retaining its computational efficiency and, under conditions on the causal graph, its axiomatic guarantees. In many settings, some causal structure between features is already known, for example, that a transcription factor causally regulates a target gene's expression. Yet existing gradient-based methods cannot exploit this structure, instead crediting a feature's downstream effects over its true upstream cause, such as ranking the target gene above the transcription factor that regulates it. Our key idea is to embed this causal structure by routing the integration path along the causal graph. The method thereby decomposes each feature's attribution into a direct effect and effects mediated through its causal descendants. We evaluate CIG in three settings of increasing realism. First, on a synthetic benchmark with a known causal model, CIG's attributions best match the ground-truth causal effect from Pearl's do-calculus, matching the accuracy of causal Shapley values at a fraction of the computational cost. Second, on simulated single-cell gene expression data with a known gene regulatory network (Dyngen), CIG most accurately recovers the transcription factors driving the simulation. Third, on real interferon-

β

stimulation data, where only an approximate causal graph is available, CIG better prioritizes the transcription factors driving the measured response than DeepSHAP and integrated gradients. Together, these results show that incorporating causal structure into gradient-based attributions recovers upstream drivers that existing methods miss, without sacrificing scalability.