Welcome to Quasimodo demo


Description

Quasimodo is an open-source commonsense knowledge base designed to get relevant properties about entities.

Abstract

Commonsense knowledge about object properties, human behavior and general concepts is crucial for robust AI applications. However, automatic acquisition of this knowledge is challenging because of sparseness and bias in online sources. This paper presents Quasimodo, a methodology and tool suite for distilling commonsense properties from non-standard web sources. We devise novel ways of tapping into search-engine query logs and QA forums, and combining the resulting candidate assertions with statistical cues from encyclopedias, books and image tags in a corroboration step. Unlike prior work on commonsense knowledge bases, Quasimodo focuses on salient properties that are typically associated with certain objects or concepts. Extensive evaluations, including extrinsic use-case studies, show that Quasimodo provides better coverage than state-of-the-art baselines with comparable quality.

Full Paper

The full paper is freely available on Arxiv. More related papers can be found on the publication page.

Data

The raw data of Quasimodo can be found on the Max Planck Institute D5's website .
Otherwise, the dataset can be explored in the Explorer section. We also added the possibility to search for specific subject, property and object in the Search section. Here are some examples of searches:

In addition, it is possible to access to more details for a given fact, such as this one.

Code

The code for running Quasimodo is available on Github. The code for running Quasimodo demo is also available on Github Github.

Pipeline

Quasimodo pipeline can be summarize by the following diagram. For more information, please refer to our paper.

Pipeline

Examples of pipeline execution for a single subject can be found on the dedicated page.