Subgoal chaining and the local minimum problem
نویسندگان
چکیده
It is well known that performing gradient descent on fixed surfaces may result in poor travel through getting stuck in local minima and other surface features. Subgoal chaining in supervised learning is a method to improve travel for neural networks by directing local variation in the surface during training. This paper shows however that linear subgoal chains such as those used in ERA are not sufficient to overcome the local minimum problem and examines non-linear subgoal chains as a possible alternative.
منابع مشابه
Microsoft Word - era978
It is well known that performing gradient descent on fixed surfaces may result in poor travel through getting stuck in local minima and other surface features. Subgoal chaining in supervised learning is a method to improve travel for neural networks by directing local variation in the surface during training. This paper shows however that linear subgoal chains such as those used in ERA are not ...
متن کاملTruthful Monadic Abstractions
In intuitionistic sequent calculi, detecting that a sequent is unprovable is often used to direct proof search. This is for instance seen in backward chaining, where an unprovable subgoal means that the proof search must backtrack. In undecidable logics, however, proof search may continue indefinitely, finding neither a proof nor a disproof of a given subgoal. In this paper we characterize a fa...
متن کاملAutomatic Logic Circuit Design with Progressive Evolution that Autonomously Gets
The progressive evolution method is a promising way to accelerate learning by making the problem size small in genetic learning. This method leads the population to the final goal by giving step-by-step learning targets, called subgoals. Giving subgoals divides a large search space into a small one, thereby accelerating the evolution. Previous implementation of progressive learning is, however,...
متن کاملType-2 Fuzzy Hybrid Expert System For Diagnosis Of Degenerative Disc Diseases
One-third of the people with an age over twenty have some signs of degenerated discs. However, in most of the patients the mere presence of degenerative discs is not a problem leading to pain, neurological compression, or other symptoms. This paper presents an interval type-2 fuzzy hybrid rule-based system to diagnose the abnormal degenerated discs where pain variables are represented by interv...
متن کاملA Metaheuristic Algorithm for the Minimum Routing Cost Spanning Tree Problem
The routing cost of a spanning tree in a weighted and connected graph is defined as the total length of paths between all pairs of vertices. The objective of the minimum routing cost spanning tree problem is to find a spanning tree such that its routing cost is minimum. This is an NP-Hard problem that we present a GRASP with path-relinking metaheuristic algorithm for it. GRASP is a multi-start ...
متن کامل