Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Go for a Walk and Arrive at the Answer: Reasoning Over Paths in Knowledge Bases using Reinforcement Learning

About

Knowledge bases (KB), both automatically and manually constructed, are often incomplete --- many valid facts can be inferred from the KB by synthesizing existing information. A popular approach to KB completion is to infer new relations by combinatory reasoning over the information found along other paths connecting a pair of entities. Given the enormous size of KBs and the exponential number of paths, previous path-based models have considered only the problem of predicting a missing relation given two entities or evaluating the truth of a proposed triple. Additionally, these methods have traditionally used random paths between fixed entity pairs or more recently learned to pick paths between them. We propose a new algorithm MINERVA, which addresses the much more difficult and practical task of answering questions where the relation is known, but only one entity. Since random walks are impractical in a setting with combinatorially many destinations from a start node, we present a neural reinforcement learning approach which learns how to navigate the graph conditioned on the input query to find predictive paths. Empirically, this approach obtains state-of-the-art results on several datasets, significantly outperforming prior methods.

Rajarshi Das, Shehzaad Dhuliawala, Manzil Zaheer, Luke Vilnis, Ishan Durugkar, Akshay Krishnamurthy, Alex Smola, Andrew McCallum• 2017

Related benchmarks

TaskDatasetResultRank
Link PredictionFB15k-237 (test)
Hits@1045.6
419
Link PredictionWN18RR (test)
Hits@1051.3
380
Link PredictionFB15k-237
MRR20.5
280
Knowledge Graph CompletionWN18RR
Hits@10.351
165
Temporal Knowledge Graph reasoningICEWS 18
Hits@100.33
60
Link PredictionUMLS
Hits@1096.8
56
Temporal Knowledge Graph reasoningICEWS 14
Hits@125.7
48
Link PredictionKinship
MRR0.72
36
Link PredictionNELL-995 (test)
MRR20.1
27
Temporal Knowledge Graph reasoningGDELT
MRR12.1
25
Showing 10 of 32 rows

Other info

Follow for update