Timezone: »

Exploring the Whole Rashomon Set of Sparse Decision Trees
Rui Xin · Chudi Zhong · Zhi Chen · Takuya Takagi · Margo Seltzer · Cynthia Rudin

Thu Dec 01 09:00 AM -- 11:00 AM (PST) @ Hall J #1001

In any given machine learning problem, there may be many models that could explain the data almost equally well. However, most learning algorithms return only one of these models, leaving practitioners with no practical way to explore alternative models that might have desirable properties beyond what could be expressed within a loss function. The Rashomon set is the set of these all almost-optimal models. Rashomon sets can be extremely complicated, particularly for highly nonlinear function classes that allow complex interaction terms, such as decision trees. We provide the first technique for completely enumerating the Rashomon set for sparse decision trees; in fact, our work provides the first complete enumeration of any Rashomon set for a non-trivial problem with a highly nonlinear discrete function class. This allows the user an unprecedented level of control over model choice among all models that are approximately equally good. We represent the Rashomon set in a specialized data structure that supports efficient querying and sampling. We show three applications of the Rashomon set: 1) it can be used to study variable importance for the set of almost-optimal trees (as opposed to a single tree), 2) the Rashomon set for accuracy enables enumeration of the Rashomon sets for balanced accuracy and F1-score, and 3) the Rashomon set for a full dataset can be used to produce Rashomon sets constructed with only subsets of the data set. Thus, we are able to examine Rashomon sets across problems with a new lens, enabling users to choose models rather than be at the mercy of an algorithm that produces only a single model.

Author Information

Rui Xin (Duke University)
Chudi Zhong (Duke University)
Zhi Chen (Duke University)
Takuya Takagi (Fujitsu Ltd.)
Margo Seltzer (University of British Columbia)

**MARGO I. SELTZER** is Canada 150 Research Chair in Computer Systems and the Cheriton Family chair in Computer Science at the University of British Columbia. Her research interests are in systems, construed quite broadly: systems for capturing and accessing data provenance, file systems, databases, transaction processing systems, storage and analysis of graph-structured data, new architectures for parallelizing execution, and systems that apply technology to problems in healthcare. She is the author of several widely-used software packages including database and transaction libraries and the 4.4BSD log-structured file system. Dr. Seltzer was a co-founder and CTO of Sleepycat Software, the makers of Berkeley DB, recipient of the 2020 ACM SIGMOD Systems Award. She serves on Advisory Council for the Canadian COVID alert app and the Computer Science and Telecommunications Board (CSTB) of the (US) National Academies. She is a past President of the USENIX Assocation and served as the USENIX representative to the Computing Research Association Board of Directors and on the Computing Community Consortium. She is a member of the National Academy of Engineering, the American Academy of Arts and Sciences, a Sloan Foundation Fellow in Computer Science, an ACM Fellow, a Bunting Fellow, and was the recipient of the 1996 Radcliffe Junior Faculty Fellowship. She is recognized as an outstanding teacher and mentor, having received the Phi Beta Kappa teaching award in 1996, the Abrahmson Teaching Award in 1999, the Capers and Marion McDonald Award for Excellence in Mentoring and Advising in 2010, and the CRA-E Undergraduate Research Mentoring Award in 2017. Professor Seltzer received an A.B. degree in Applied Mathematics from Harvard/Radcliffe College and a Ph. D. in Computer Science from the University of California, Berkeley.

Cynthia Rudin (Duke)

More from the Same Authors