PDF4PRO ⚡AMP

Modern search engine that looking for books and documents around the web

Example: tourism industry

Anchors: High Precision Model-Agnostic Explanations

Anchors: High- Precision Model-Agnostic ExplanationsMarco Tulio RibeiroUniversity of SinghUniversity of California, GuestrinUniversity of introduce a novel Model-Agnostic system that explains thebehavior of complex models with high- Precision rules calledanchors, representing local, sufficient conditions for predic-tions. We propose an algorithm to efficiently compute theseexplanations for any black-box model with high-probabilityguarantees. We demonstrate the flexibility of anchors by ex-plaining a myriad of different models for different domainsand tasks. In a user study, we show that anchors enable usersto predict how a model would behave on unseen instanceswith less effort and higher Precision , as compared to existinglinear Explanations or no machine learning models such as deep neuralnetworks have been shown to be highly accurate for manyapplicat

In this paper, we introduce novel model-agnostic expla-nations based on if-then rules, which we call anchors. An anchor explanation is a rule that sufficiently “anchors” the prediction locally – such that changes to the rest of the feature values of …

Fullscreen Download
Loading..

Tags:

  Paper, Anchor

Information

Domain:

Source:

Link to this page:

Please notify us if you found a problem with this document:

Spam in document Broken preview Other abuse

Transcription of Anchors: High Precision Model-Agnostic Explanations

Documents from same domain

Evaluating&improving fault localization techniques

homes.cs.washington.edu

Evaluating&improving fault localization techniques Technical report UW-CSE-16-08-03 August 2016; revised February 2017 ... The standard technique for evaluating fault localization, described in section II-A, handles defects that consist of a change to one …

  Technique, Improving, Evaluating, Fault, Localization, Evaluating amp improving fault localization techniques

Game Theory, Alive - University of Washington

homes.cs.washington.edu

Acknowledgements We are grateful to Alan Hammond, Yun Long, G abor Pete, and Peter Ralph for scribing early drafts of some of the chapters in this book from lectures by Yuval

  Games, Theory, Alive, Game theory

Chapter 3 Fundamental Laws - University of Washington

homes.cs.washington.edu

3.3. Little’s Law Hence: 43 Little’s Law: N = XR That is, the average number of requests in a system is equal to the pro- duct of the throughput of that system and the average time spent in that

  Little, Little s law

Sloth: Locating Sites for Repetitive Edits with Lazy ...

homes.cs.washington.edu

Sloth: Locating Sites for Repetitive Edits with Lazy Concrete Pattern Matching on Trees Conference’17, July 2017, Washington, DC, USA matches on any identifier but ignores what construct the identifier

  Sloth

Identification and control of a pneumatic robot

homes.cs.washington.edu

Identification and control of a pneumatic robot Emanuel Todorov 1,ChunyanHu, Alex Simpkins1 and Javier Movellan2 1Applied Mathematics and Computer Science and Engineering, University of Washington 2 Institute for Neural Computation, University of California San Diego Abstract—Pneumatic actuators have a number of advantages over electric motors, including strength-to-weight ratio, tunable

  Control, Identification, Identification and control of a

Particle Filters - University of Washington

homes.cs.washington.edu

Example 3: Example Particle Distributions [Grisetti, Stachniss, Burgard, T-RO2006] Particles generated from the approximately optimal proposal distribution. If using the standard motion model, in all three cases the particle set would have been similar to (c). "

  Particles, Filter, Particle filters

Policy Gradient Methods for Reinforcement Learning with ...

homes.cs.washington.edu

Williams’s (1988, 1992) REINFORCE algorithm also flnds an unbiased estimate of the gradient, but without the assistance of a learned value function. REINFORCE learns much more slowly than RL methods using value functions and has received relatively little attention. Learning a value function and using it to reduce the variance

  Methods, Learning, Reinforcement, Derating, Gradient methods for reinforcement learning

Jacobian methods for inverse kinematics and planning

homes.cs.washington.edu

Operating Principle: - Project difference vector Dx on those dimensions q which can reduce it ... Equality constraints Such constraints restrict the state to a manifold. If the simple push-towards-the-goal action projected on the manifold always gets us closer to the goal, then the problem is …

  Principles, Differences, Equality

Optimal Control Theory - homes.cs.washington.edu

homes.cs.washington.edu

Equations (1, 3, 4) generalize to the stochastic case in the same way as equation (2) does. An optimal control problem with discrete states and actions and probabilistic state transitions is called a Markov decision process (MDP).

  Control, Theory, Optimal, Stochastic, Optimal control theory

Chapter 14 Proposed Systems - University of Washington

homes.cs.washington.edu

In the late 1960s TSO was being developed as a timesharing subsys- tem for IBM’s batch-oriented MVT operating system. During final design and initial implementation of the final system, an earlier prototype was measured in a test environment, and a queueing network model was

  Subsys