{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/how-important-is-a-neuron","title":"How Important Is a Neuron?","arxiv_id":"1805.12233","date":"2018-05-30","proceeding":null,"authors":["Kedar Dhamdhere","Mukund Sundararajan","Qiqi Yan"],"abstract":"The problem of attributing a deep network's prediction to its\n\\emph{input/base} features is well-studied. We introduce the notion of\n\\emph{conductance} to extend the notion of attribution to the understanding the\nimportance of \\emph{hidden} units.\n  Informally, the conductance of a hidden unit of a deep network is the\n\\emph{flow} of attribution via this hidden unit. We use conductance to\nunderstand the importance of a hidden unit to the prediction for a specific\ninput, or over a set of inputs. We evaluate the effectiveness of conductance in\nmultiple ways, including theoretical properties, ablation studies, and a\nfeature selection task. The empirical evaluations are done using the Inception\nnetwork over ImageNet data, and a sentiment analysis network over reviews. In\nboth cases, we demonstrate the effectiveness of conductance in identifying\ninteresting insights about the internal workings of these networks.","url_abs":"http://arxiv.org/abs/1805.12233v1","url_pdf":"http://arxiv.org/pdf/1805.12233v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"how-important-is-a-neuron","repo_url":"https://github.com/feiwu77777/Leveraging-Model-Interpretability-and-Stability-to-increase-Model-Robustness","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"how-important-is-a-neuron","repo_url":"https://github.com/pytorch/captum","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"BSD-3-Clause"}}],"tasks":[{"task_slug":"sentiment-analysis","task_name":"Sentiment Analysis"},{"task_slug":"feature-selection","task_name":"feature selection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1805.12233","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}