Newest papers

Causal Integrated Gradients: Scalable Feature Attribution Along Causal Paths

Joran Michiels, Anis Ismail, Maarten De Vos, Alejandro Sifrim · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.22126 · cited by 0 works (OpenAlex) · open-access PDF

Indonesian

Metode atribusi gradient yang didorong oleh grass-cell yang ada adalah metode atribusi grass-cell yang memperhitungkan ketergantungan data yang tidak meluas ke dimensi-dimensi tinggi. Kami mengusulkan gradient transcendent (CIG) dan gradient transcendent (CEG), yang memperluas gradient integrasi untuk menggabungkan struktur kausal sebelumnya di antara fitur input sambil mempertahankan efisiensi komputasi dan, di bawah kondisi pada grafik kausal, jaminan aksiomatiknya. Dalam banyak pengaturan, beberapa struktur kausal antara fitur sudah diketahui, misalnya, bahwa faktor transkripsi secara kausal mengatur ekspresi gen target. Namun metode berbasis gradient yang ada tidak dapat memanfaatkan struktur ini, alih-alih mempercayai efek downstream dari fitur pada penyebabnya yang sebenarnya, seperti peringkat gen target di atas faktor tran

Machine translation (Meta M2M100); the English original follows.

Abstract

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.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Causal model, Causal structure, Attribution, Scalability, Artificial intelligence

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange