Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Properly Learning Decision Trees in almost Polynomial Time

Journal Article · · Journal of the Association for Computing Machinery
DOI:https://doi.org/10.1145/3561047· OSTI ID:2421027

We give annO(log logn)-time membership query algorithm for properly and agnostically learning decision trees under the uniform distribution over { ± 1}n. Even in the realizable setting, the previous fastest runtime wasnO(logn), a consequence of a classic algorithm of Ehrenfeucht and Haussler.

Our algorithm shares similarities with practical heuristics for learning decision trees, which we augment with additional ideas to circumvent known lower bounds against these heuristics. To analyze our algorithm, we prove a new structural result for decision trees that strengthens a theorem of O’Donnell, Saks, Schramm, and Servedio. While the OSSS theorem says that every decision tree has an influential variable, we show how every decision tree can be “pruned” so thateveryvariable in the resulting tree is influential.

Research Organization:
Stanford Univ., CA (United States)
Sponsoring Organization:
USDOE Office of Science (SC)
DOE Contract Number:
SC0019205
OSTI ID:
2421027
Journal Information:
Journal of the Association for Computing Machinery, Journal Name: Journal of the Association for Computing Machinery Journal Issue: 6 Vol. 69; ISSN 0004-5411
Publisher:
Association for Computing Machinery (ACM)
Country of Publication:
United States
Language:
English

References (19)

Learning functions of k relevant variables journal November 2004
Learning Decision Trees Using the Fourier Spectrum journal December 1993
Learning and Smoothed Analysis conference October 2009
Constructing optimal binary decision trees is NP-complete journal May 1976
Finding Correlations in Subquadratic Time, with Applications to Learning Parities and the Closest Pair Problem journal May 2015
Approximating Optimal Binary Decision Trees journal April 2011
Learning Monotone Decision Trees in Polynomial Time journal January 2007
Finding Small Equivalent Decision Trees is hard journal June 2000
Rank-r decision trees are a subclass of r-decision lists journal June 1992
Minimization of decision trees is hard to approximate journal May 2008
Beyond the low-degree algorithm: mixtures of subcubes and their applications conference June 2019
Selection of relevant features and examples in machine learning journal December 1997
Decision tree approximations of Boolean functions journal January 2002
Agnostically learning decision trees conference May 2008
Weakly learning DNF and characterizing statistical query learning using Fourier analysis conference January 1994
Sharp phase transition for the random-cluster and Potts models via decision trees journal January 2019
Learning decision trees from random examples journal September 1989
Learning kμ decision trees on the uniform distribution conference January 1993
Lower Bounds on Learning Decision Lists and Trees journal May 1996

Similar Records

Inferring hierarchical clustering structures by deterministic annealing
Conference · Mon Dec 30 23:00:00 EST 1996 · OSTI ID:421314

Polynomial time approximation schemes for Euclidean TSP and other geometric problems
Conference · Mon Dec 30 23:00:00 EST 1996 · OSTI ID:457631

A polynomial time primal network simplex algorithm for minimum cost flows
Conference · Mon Dec 30 23:00:00 EST 1996 · OSTI ID:416832

Related Subjects