Tsetlin Machine (TM) is a logic-based machine learning approach with the crucial advantages of being transparent and hardware-friendly. While TMs match or surpass deep learning accuracy for an increasing number of applications, large clause pools tend to produce clauses with many literals (long clauses). As such, they become less interpretable. Further, longer clauses increase the switching activity of the clause logic in hardware, consuming more power. This paper introduces a novel variant of TM learning -- Clause Size Constrained TMs (CSC-TMs) -- where one can set a soft constraint on the clause size. As soon as a clause includes more literals than the constraint allows, it starts expelling literals. Accordingly, oversized clauses only appear transiently. To evaluate CSC-TM, we conduct classification, clustering, and regression experiments on tabular data, natural language text, images, and board games. Our results show that CSC-TM maintains accuracy with up to 80 times fewer literals. Indeed, the accuracy increases with shorter clauses for TREC and BBC Sports. After the accuracy peaks, it drops gracefully as the clause size approaches one literal. We finally analyze CSC-TM power consumption and derive new convergence properties.

Abstract Tsetlin Machines (TM) use finite state machines for learning and propositional logic to represent patterns. The resulting pattern recognition approach captures information in the form of conjunctive clauses, thus facilitating human interpretation. In this work, we propose a TM‐based approach to three common natural language processing (NLP) tasks, namely, sentiment analysis, semantic relation categorization and identifying entities in multi‐turn dialogues. By performing frequent itemset mining on the TM‐produced patterns, we show that we can obtain a global and a local interpretation of the learning, one that mimics existing rule‐sets or lexicons. Further, we also establish that our TM based approach does not compromise on accuracy in the quest for interpretability, via comparison with some widely used machine learning techniques. Finally, we introduce the idea of a relational TM, which uses a logic‐based framework to further extend the interpretability.

Abstract Tsetlin machines (TMs) are a pattern recognition approach that uses finite state machines for learning and propositional logic to represent patterns. In addition to being natively interpretable, they have provided competitive accuracy for various tasks. In this paper, we increase the computing power of TMs by proposing a first-order logic-based framework with Herbrand semantics. The resulting TM isrelationaland can take advantage of logical structures appearing in natural language, to learn rules that represent how actions and consequences are related in the real world. The outcome is a logic program of Horn clauses, bringing in a structured view of unstructured data. In closed-domain question-answering, the first-order representation produces 10 × more compact KBs, along with an increase in answering accuracy from 94.83%to 99.48%. The approach is further robust towards erroneous, missing, and superfluous information, distilling the aspects of a text that are important for real-world understanding

Using logical clauses to represent patterns, Tsetlin Machine (TM) have recently obtained competitive performance in terms of accuracy, memory footprint, energy, and learning speed on several benchmarks. Each TM clause votes for or against a particular class, with classification resolved using a majority vote. While the evaluation of clauses is fast, being based on binary operators, the voting makes it necessary to synchronize the clause evaluation, impeding parallelization. In this paper, we propose a novel scheme for desynchronizing the evaluation of clauses, eliminating the voting bottleneck. In brief, every clause runs in its own thread for massive native parallelism. For each training example, we keep track of the class votes obtained from the clauses in local voting tallies. The local voting tallies allow us to detach the processing of each clause from the rest of the clauses, supporting decentralized learning. This means that the TM most of the time will operate on outdated voting tallies. We evaluated the proposed parallelization across diverse learning tasks and it turns out that our decentralized TM learning algorithm copes well with working on outdated data, resulting in no significant loss in learning accuracy. Furthermore, we show that the approach provides up to 50 times faster learning. Finally, learning time is almost constant for reasonable clause amounts (employing from 20 to 7,000 clauses on a Tesla V100 GPU). For sufficiently large clause numbers, computation time increases approximately proportionally. Our parallel and asynchronous architecture thus allows processing of more massive datasets and operating with more clauses for higher accuracy.

In this paper we have proposed a linguistically informed recursive neural network architecture for automatic extraction of cause-effect relations from text.These relations can be expressed in arbitrarily complex ways.The architecture uses word level embeddings and other linguistic features to detect causal events and their effects mentioned within a sentence.The extracted events and their relations are used to build a causal-graph after clustering and appropriate generalization, which is then used for predictive purposes.We have evaluated the performance of the proposed extraction model with respect to two baseline systems,one a rule-based classifier, and the other a conditional random field (CRF) based supervised model.We have also compared our results with related work reported in the past by other authors on SEMEVAL data set, and found that the proposed bidirectional LSTM model enhanced with an additional linguistic layer performs better.We have also worked extensively on creating new annotated datasets from publicly available data, which we are willing to share with the community.

In this paper, we have proposed natural language processing and deep learning based techniques for the automatic extraction and curation of occupational health and safety related information from safety-related articles. Such articles typically contain details of the organizations that have been cited for violating the health and safety regulations, safety-related issues and incidents, the location of the incident, and finally details of the penalties incurred. We have done experiments with a collection of 5400 related articles. The end-product of our work is an occupational risk-register that contains details of safety incidents across geographies and time. This register can be further utilized for analytical and reporting purposes. Such information is extremely valuable to industries which see a high occurrence of occupational injuries.

News articles from different sources regularly report crime incidents that contain details of crime, information about accused entities, details of the investigation process and finally details of judgement. In this paper, we have proposed natural language processing techniques for extraction and curation of crime-related information from digitally published News articles. We have leveraged computational linguistics based methods to analyse crime related News documents to extract different crime related entities and events. This includes name of the criminal, name of the victim, nature of crime, geographic location, date and time, and action taken against the criminal. We have also proposed a semi-supervised learning technique to learn different categories of crime events from the News documents. This helps in continuous evolution of the crime dictionaries. Thus the proposed methods are not restricted to detecting known crimes only but contribute actively towards maintaining an updated crime dictionary. We have done experiments with a collection of 3000 crime-reporting News articles. The end-product of our experiments is a crime-register that contains details of crime committed across geographies and time. This register can be further utilized for analytical and reporting purposes.