Building Concise Logical Patterns by Constraining Tsetlin Machine Clause Size

Authors: K. Darshana Abeyrathna, Ahmed A. O. Abouzeid, Bimal Bhattarai, Charul Giri, Sondre Glimsdal, Ole-Christoffer Granmo, Lei Jiao, Rupsa Saha, Jivitesh Sharma, Svein A. Tunheim, Xuan Zhang

IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 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.
Researcher Affiliation Collaboration 1Centre for Artificial Intelligence Research (CAIR), University of Agder, Grimstad, Norway 2Norwegian Research Centre (NORCE), Grimstad, Norway 3DNV, Oslo, Norway
Pseudocode No No structured pseudocode or algorithm blocks were found. Table 1 presents feedback rules, but it is not pseudocode.
Open Source Code No The paper refers to a general TM learning resource (https://cair.github.io/ijcai2023_clause_rationing.html) but does not explicitly state that the source code for the CSC-TM methodology described in this paper is released or provided.
Open Datasets Yes We evaluate CSC-TM on five NLP datasets: BBC sports [Greene and Cunningham, 2006], R8 [Debole and Sebastiani, 2005], TREC-6 [Chang et al., 2002], Sem Eval 2010 Semantic Relations [Hendrickx et al., 2009], and ACL Internet Movie Database (IMDb) [Maas et al., 2011]. ... We evaluate our approach on two image datasets: MNIST and CIFAR-2... We use the Energy Performance dataset to evaluate regression performance based on [Abeyrathna et al., 2020c].
Dataset Splits No The paper mentions using well-known datasets like MNIST but does not explicitly provide the training/test/validation dataset splits (e.g., percentages or sample counts) needed for reproduction.
Hardware Specification Yes The experiments use a CUDA implementation of CSC-TM and runs on Intel Xeon Platinum 8168 CPU at 2.70 GHz and a Nvidia DGX-2 with Tesla V100 GPU.
Software Dependencies No The paper mentions 'CUDA implementation' but does not provide specific version numbers for software dependencies or libraries.
Experiment Setup Yes For TM hyperparameters in BBC Sports, TREC, and R8, we use 8000 clauses, a voting margin T of 100, and specificity s of 10.0. For hyperparameters, we adopt 8000 clauses per class, a voting margin T as 10000, and specificity s as 5.0 in the MNIST experiments. For CIFAR-2, the number of clauses is 8000, T is 6000, and s is 10.0.