Is it a Fruit, an Apple or a Granny Smith? Predicting the Basic Level in a Concept Hierarchy

arXiv.org Artificial Intelligence

Is it a Fruit, an Apple or a Granny Smith? Abstract The "basic level", according to experiments in cognitive psychology, is the level of abstraction in a hierarchy of concepts at which humans perform tasks quicker and with greater accuracy than at other levels. We argue that applications that use concept hierarchies - such as knowledge graphs, ontologies or taxonomies - could significantly improve their user interfaces if they'knew' which concepts are the basic level concepts. This paper examines to what extent the basic level can be learned from data. We test the utility of three types of concept features, that were inspired by the basic level theory: lexical features, structural features and frequency features. We evaluate our approach on WordNet, and create a training set of manually labelled examples that includes concepts from different domains. Our findings include that the basic level concepts can be accurately identified within one domain. Concepts that are difficult to label for humans are also harder to classify automatically. Our experiments provide insight into how classification performance across domains could be improved, which is necessary for identification of basic level concepts on a larger scale. 1 Introduction One of the ongoing challenges in Artificial Intelligence is to explicitly describe the world in ways that machines can process. This has resulted in taxonomies, thesauri, ontologies and more recently knowledge graphs. While these various knowledge organization systems (KOSs) may use different formal languages, they all share similar underlying data representations.


Impact of Lexical Filtering on Overall Opinion Polarity Identification

AAAI Conferences

One approach to assessing overall opinion polarity (OvOP) of reviews, a concept defined in this paper, is the use of supervised machine learning mechanisms. In this paper, the impact of lexical filtering, applied to reviews, on the accuracy of two statistical classifiers (Naive Bayes and Markov Model) with respect to OvOP identification is observed. Two kinds of lexical filters, one based on hypernymy as provided by Word-Net (Fellbaum 1998), and one handcrafted filter based on part-of-speech (POS) tags, are evaluated. A ranking criterion based on a function of the probability of having positive or negative polarity is introduced and verified as being capable of achieving 100% accuracy with 10% recall. Movie reviews are used for training and evaluation of each statistical classifier, achieving 80% accuracy.


Persian Wordnet Construction using Supervised Learning

arXiv.org Machine Learning

This paper presents an automated supervised method for Persian wordnet construction. Using a Persian corpus and a bi-lingual dictionary, the initial links between Persian words and Princeton WordNet synsets have been generated. These links will be discriminated later as correct or incorrect by employing seven features in a trained classification system. The whole method is just a classification system, which has been trained on a train set containing FarsNet as a set of correct instances. State of the art results on the automatically derived Persian wordnet is achieved. The resulted wordnet with a precision of 91.18% includes more than 16,000 words and 22,000 synsets.


Learning Syntactic Patterns for Automatic Hypernym Discovery

Neural Information Processing Systems

Semantic taxonomies such as WordNet provide a rich source of knowledge fornatural language processing applications, but are expensive to build, maintain, and extend. Motivated by the problem of automatically constructing and extending such taxonomies, in this paper we present a new algorithm for automatically learning hypernym (is-a) relations from text. Our method generalizes earlier work that had relied on using small numbers of handcrafted regular expression patterns to identify hypernym pairs.Using "dependency path" features extracted from parse trees, we introduce a general-purpose formalization and generalization of these patterns. Given a training set of text containing known hypernym pairs, our algorithm automatically extracts useful dependency paths and applies them to new corpora to identify novel pairs. On our evaluation task (determining whethertwo nouns in a news article participate in a hypernym relationship), our automatically extracted database of hypernyms attains both higher precision and higher recall than WordNet.


An Automatic Method for Generating Sense Tagged Corpora

AAAI Conferences

The following resources have been used in developing and testing our method. WordNet 1 is a Machine Readable Dictionary developed at Princeton University by a group led by George Miller (Miller 1995), (Fellbaum 1998).