Open Information Extraction

• Open IE 5: The latest software for Open Information Extraction. It additionally includes Numerical Open IE and advanced conjunction processing.
• Open IE Demo: A demo for querying the database of Open IE triples extracted from the Web.
• OREO: A rapidly retargetable system for extracting information in a new domain.
• Open IE 4: The 4th generation software for Open Information Extraction that combines SRLIE and RelNoun.
• OLLIE: The 3rd generation software for Open Information Extraction. (relevant paper)
• ReVerb: The 2nd generation software for Open Information Extraction. (relevant paper)
• Rel-grams: A corpus of relational n-grams -- relational co-occurences in news stories. (relevant paper)
• LDA-SP: A demo of selectional preferences (types for each argument) for Open IE relational phrases. (relevant paper)
• Functional Relations: A list of 10K relational phrases that satisfy the functionality property. (relevant paper)
• Entity Linking: Three million ReVerb extractions with arguments linked to FreeBase. (relevant paper)

Crowdsourcing

• Sprout: A software tool to aid requesters in improving the task design. (relevant paper)
• Octopus: A software optimizing the cost-quality-completion time tradeoff for a crowdsourcing task. (relevant paper)
• Re-Active Learning: A software for active learning with relabeling to train ML classifiers over crowdsourced platforms. (relevant papers 1, 2)
• High Skew ML: A software for active learning for high-skew domains for training ML classifiers over crowdsourced platforms. (relevant paper)
• Guru: A software to optimize testing of workers for a crowdsourcing task. (relevant paper)
• Juggler: A software for task routing of multiple tasks to multiple workers simultaneously. (relevant paper)
• Deluge: A software for control of multi-label classification tasks. (relevant paper)

Inference in NLP

• KBI: An implementation of tensor factorization> and matrix factorization models for KB inference.
• KGLR: A high precision dataset of Open IE inference rules. (relevant paper)
• TRINE: An extraction and inference engine to extract and predict recurrent events. (relevant paper)

NLP over Twitter

• TwitCal: A demo of popular past, present, and future events based on Twitter stream. (relevant paper)
• Twitter NLP Pipeline: Software for POS Tagging, NP Chunking and NER over Twitter data. (relevant paper)

Multi-Document Summarization

• GFlow: A software and demo for summarizing a cluster of related news articles. (relevant paper)
• Summa: A demo for hierarchically summarizing a large cluster of related news articles. (relevant paper)

Probabilistic Planning

• OGA-UCT: An extension of ASAP-UCT for solving large finite-horizon MDPs using state-action pair abstractions. (relevant paper)
• ASAP-UCT: A software for solving large finite-horizon MDPs using state-action pair abstractions. (relevant paper)
• Glutton: A software for solving large finite-horizon MDPs with high branching factors (relevant paper)
• Gourmand: An improvement over Glutton -- a software for solving large finite-horizon MDPs with high branching factors (relevant paper)
• ReTrASE: A probabilistic planner to solve large goal-oriented MDPs. The solver is available on request. (relevant paper)
• CoMDP Solver: A concurrent MDP solver to solve planning under uncertainty problems that require executing more than one action at a time step. (relevant paper)

Probabilistic Graphical Models

• BV-MCMC: Block-Value Symmetries in MCMC (relevant paper)
• VV-MCMC: Variable-Value (Non-Count) Symmetries in MCMC (relevant paper)
• Con-MCMC: Contextual Symmetries in MCMC (relevant paper)

Computer Vision

• C2F-CV: Coarse-to-fine algorithms for faster inference in computer vision (relevant paper)

Machine Translation

• PanDictionary: A massively multilingual dictionary of words and phrases with over 1,000 languages. The dictionary is available on request. (relevant paper)