Toward an Architecture for Never-Ending Language Learning

By Andrew Carlson, Justin Betteridge, Bryan Kisiel, Burr Settles, Estevam R. Hruschka, Tom Michael Mitchell. 2010

2,255 citations recorded by the source corpus.

We consider here the problem of building a never-ending language learner; that is, an intelligent computer agent that runs forever and that each day must (1) extract, or read, information from the web to populate a growing structured knowledge base, and (2) learn to perform this task better than on the previous day. In particular, we propose an approach and a set of design principles for such an agent, describe a partial implementation of such a system that has already learned to extract a knowledge base containing over 242,000 beliefs with an estimated precision of 74% after running for 67 days, and discuss lessons learned from this preliminary attempt to build a never-ending learning agent.

Subfields: Artificial Intelligence, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

Citations in the corpus

Long Short-Term Memory-Networks for Machine Reading (2016, background): “, 2011), knowledge base population and construction (Ji and Grishman, 2011; Niu et al., 2012; Carlson et al., 2010), ontology learning (Poon and Domingos, 2010b), textual entailment (Dagan et al.”
Semi-Supervised Classification with Graph Convolutional Networks (2016, background): “NELL NELL is a dataset extracted from the knowledge graph introduced in (Carlson et al., 2010).”
Experience Grounds Language (2020, background): “With few exceptions (Carlson et al., 2010), machine learning models have been confined to IID datasets that lack the structure”
Semantic Parsing on Freebase from Question-Answer Pairs (2013, background): “Sev- 8Oracle score is the fraction of examples for which D̃(x) contains any derivation with the correct denotation. eral works perform relation extraction using distant supervision from a knowledge base (Riedel et al., 2010; Carlson et al., 2010; Hoffmann et al., 2011; Surdeanu et al., 2012).”
Open-Domain Conversational Agents: Current Progress, Open Problems, and Future Directions (2020, background): “The general challenges of never-ending learning (Mitchell et al., 2018; Carlson et al., 2010) and avoiding catastrophic forgetting (French, 1999; Kirkpatrick et al.”
Key-Value Memory Networks for Directly Reading Documents (2016, background): “Due to the sparsity of KB data, however, the main challenge shifts from ﬁnding answers to developing efﬁcient information extraction methods to populate KBs automatically (Craven et al. , 2000; Carlson et al. , 2010) – which is not an easy problem.”
Learning from Dialogue after Deployment: Feed Yourself, Chatbot! (2019, background): “, 2013) and never-ending (language) learning (Carlson et al., 2010) are related to the topics dis-”
Revisiting Semi-Supervised Learning with Graph Embeddings (2016, methodology): “We ﬁrst remove relations that are not populated in NELL, including “generalizations”, “haswikipediaurl”, and “atdate”.”
BlenderBot 3: a deployed conversational agent that continually learns to responsibly engage (2022, background): “…domain, there is also a rich body of work studying the improvement of models from deployment, including never-endinglearning from language data (Carlson et al., 2010), improving web search (Agichtein et al., 2006), the Dynabench system which evaluates a number of NLP tasks (Kiela et al.,…”
Mass-Editing Memory in a Transformer (2022, background): “Pioneering hand-curated efforts (Lenat, 1995; Miller, 1995) have been followed by web-powered knowledge graphs (Auer et al., 2007; Bollacker et al., 2007; Suchanek et al., 2007; Havasi et al., 2007; Carlson et al., 2010; Dong et al., 2014; Vrandečić & Krötzsch, 2014; Bosselut et al., 2019) that extr”
From captions to visual concepts and back (2014, background): “This training captures commonsense knowledge about the world through language statistics [3].”
COMET: Commonsense Transformers for Automatic Knowledge Graph Construction (2019, background): “…extracting semi-structured text (Suchanek et al., 2007; Hoffart et al., 2013; Auer et al., 2007; Bollacker et al., 2008) and unstructured text (Dong et al., 2014; Carlson et al., 2010; Nakashole et al., 2011, 2012; Niu, 2012) into relational schemas that can be queried for downstream applications.”
Modeling Biological Processes for Reading Comprehension (2014, background): “Work in this ﬁeld has focused mostly on macro-reading, i.e., processing large text collections and extracting knowledge bases of facts (Et-zioni et al., 2006; Carlson et al., 2010; Fader et al., 2011).”
Gradient Episodic Memory for Continual Learning (2017, background): “This learning setting led to implementations [Carlson et al., 2010, Ruvolo and Eaton, 2013, Ring, 1997], and theoretical investigations [Baxter, 2000, Balcan et al., 2015, Pentina and Urner, 2016], although the latter ones have been restricted to linear models.”
ConTinTin: Continual Learning from Task Instructions (2022, background): “…work also focuses on low-resource continual learning (Solomonoff, 1989), sequential learning (McCloskey and Cohen, 1989), and never-ending learning (Carlson et al., 2010). learning; in contrast, our learning problem belongs to TCL while each task in our formulation is expressed by instructions…”