no code implementations • LREC 2022 • Bimal Bhattarai, Ole-Christoffer Granmo, Lei Jiao
Recent advancements in natural language processing (NLP) have reshaped the industry, with powerful language models such as GPT-3 achieving superhuman performance on various tasks.
Ranked #1 on Fake News Detection on PolitiFact
no code implementations • 26 Oct 2023 • Rupsa Saha, Vladimir I. Zadorozhny, Ole-Christoffer Granmo
We propose a novel way of assessing and fusing noisy dynamic data using a Tsetlin Machine.
no code implementations • 23 Oct 2023 • Daniel Biermann, Fabrizio Palumbo, Morten Goodwin, Ole-Christoffer Granmo
As far as we are aware, no model uses the sequence length reduction step as an additional opportunity to tune the models performance.
no code implementations • 17 Oct 2023 • Bimal Bhattarai, Ole-Christoffer Granmo, Lei Jiao, Per-Arne Andersen, Svein Anders Tunheim, Rishad Shafik, Alex Yakovlev
In brief, the TA of each clause literal has both an absorbing Exclude- and an absorbing Include state, making the learning scheme absorbing instead of ergodic.
no code implementations • 3 Oct 2023 • Mohamed-Bachir Belaid, Jivitesh Sharma, Lei Jiao, Ole-Christoffer Granmo, Per-Arne Andersen, Anis Yazidi
Tsetlin Machines (TMs) have garnered increasing interest for their ability to learn concepts via propositional formulas and their proven efficiency across various application domains.
1 code implementation • 12 Sep 2023 • Ole-Christoffer Granmo, Per-Arne Andersen, Lei Jiao, Xuan Zhang, Christian Blakely, Tor Tveit
A set of variables is the Markov blanket of a random variable if it contains all the information needed for predicting the variable.
1 code implementation • 9 Sep 2023 • Ole-Christoffer Granmo
Tsetlin Machines (TMs) provide a fundamental shift from arithmetic-based to logic-based machine learning.
no code implementations • 1 Jun 2023 • Samuel Prescott, Adrian Wheeldon, Rishad Shafik, Tousif Rahman, Alex Yakovlev, Ole-Christoffer Granmo
We present use cases for online learning using the proposed infrastructure and demonstrate the energy/performance/accuracy trade-offs.
no code implementations • 19 May 2023 • Rishad Shafik, Tousif Rahman, Adrian Wheeldon, Ole-Christoffer Granmo, Alex Yakovlev
Our analyses provides the first insights into conflicting design tradeoffs involved in energy-efficient and interpretable decision models for this new artificial intelligence hardware architecture.
1 code implementation • 25 Mar 2023 • Emilia Przybysz, Bimal Bhattarai, Cosimo Persia, Ana Ozaki, Ole-Christoffer Granmo, Jivitesh Sharma
Then, we show the correctness of our encoding and provide results for the properties: adversarial robustness, equivalence, and similarity of TsMs.
no code implementations • 20 Jan 2023 • Jinbao Zhang, Xuan Zhang, Lei Jiao, Ole-Christoffer Granmo, Yongjun Qian, Fan Pan
In this study, we develop a Tsetlin machine (TM) based architecture for premature ventricular contraction (PVC) identification by analysing long-term ECG signals.
no code implementations • 19 Jan 2023 • K. Darshana Abeyrathna, Ahmed Abdulrahem Othman Abouzeid, Bimal Bhattarai, Charul Giri, Sondre Glimsdal, Ole-Christoffer Granmo, Lei Jiao, Rupsa Saha, Jivitesh Sharma, Svein Anders Tunheim, Xuan Zhang
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.
1 code implementation • 2 Jan 2023 • Bimal Bhattarai, Ole-Christoffer Granmo, Lei Jiao, Rohan Yadav, Jivitesh Sharma
We also visualize word clusters in vector space, demonstrating how our logical embedding co-locate similar words.
no code implementations • 27 Dec 2022 • Jivitesh Sharma, Ole-Christoffer Granmo, Lei Jiao
Tsetlin Machine (TM) has been gaining popularity as an inherently interpretable machine leaning method that is able to achieve promising performance with low computational complexity on a variety of applications.
no code implementations • 3 Oct 2022 • Per-Arne Andersen, Ole-Christoffer Granmo, Morten Goodwin
We show that the DVQN algorithm is a promising approach for identifying initiation and termination conditions for option-based reinforcement learning.
1 code implementation • 3 Oct 2022 • Per-Arne Andersen, Morten Goodwin, Ole-Christoffer Granmo
CaiRL also presents the first reinforcement learning toolkit with a built-in JVM and Flash support for running legacy flash games for reinforcement learning research.
no code implementations • 3 Oct 2022 • Per-Arne Andersen, Morten Goodwin, Ole-Christoffer Granmo
There are two approaches, model-based and model-free reinforcement learning, that show concrete results in several disciplines.
no code implementations • 30 Aug 2022 • Ajay Vishwanath, Einar Duenger Bøhn, Ole-Christoffer Granmo, Charl Maree, Christian Omlin
Using modern day AI techniques, such as affinity-based reinforcement learning and explainable AI, we motivate the implementation of virtuous agents that play such role-playing games, and the examination of their decisions through a virtue ethical lens.
1 code implementation • 23 Mar 2022 • Ahmed Abouzeid, Ole-Christoffer Granmo, Christian Webersik, Morten Goodwin
We further propose a generic misinformation mitigation algorithm that is robust to different social networks' misinformation statistics, allowing a promising impact in real-world scenarios.
no code implementations • 8 Mar 2022 • Charul Giri, Ole-Christoffer Granmo, Herke van Hoof, Christian D. Blakely
Hex is a turn-based two-player connection game with a high branching factor, making the game arbitrarily complex with increasing board sizes.
1 code implementation • 4 Feb 2022 • Raihan Seraj, Jivitesh Sharma, Ole-Christoffer Granmo
This paper introduces an interpretable contextual bandit algorithm using Tsetlin Machines, which solves complex pattern recognition tasks using propositional logic.
2 code implementations • 17 Sep 2021 • Lei Jiao, Xuan Zhang, Ole-Christoffer Granmo
The analyses on AND and OR operators, together with the previously analysed 1-bit and XOR operations, complete the convergence analyses on basic operators in Boolean algebra.
5 code implementations • 17 Aug 2021 • Sondre Glimsdal, Ole-Christoffer Granmo
While TM and CoTM accuracy is similar when using more than $1$K clauses per class, CoTM reaches peak accuracy $3\times$ faster on MNIST with $8$K clauses.
6 code implementations • 30 May 2021 • Jivitesh Sharma, Rohan Yadav, Ole-Christoffer Granmo, Lei Jiao
In this article, we introduce a novel variant of the Tsetlin machine (TM) that randomly drops clauses, the key learning elements of a TM.
5 code implementations • LREC 2022 • Bimal Bhattarai, Ole-Christoffer Granmo, Lei Jiao
The proliferation of fake news, i. e., news intentionally spread for misinformation, poses a threat to individuals and society.
5 code implementations • 10 May 2021 • Bimal Bhattarai, Ole-Christoffer Granmo, Lei Jiao
Our approach encodes a description of the novel documents using the linguistic patterns captured by TM clauses.
5 code implementations • EMNLP (BlackboxNLP) 2021 • Rohan Kumar Yadav, Lei Jiao, Ole-Christoffer Granmo, Morten Goodwin
The approach significantly enhances the performance and interpretability of TM.
Ranked #8 on Text Classification on R52
5 code implementations • 22 Feb 2021 • Rupsa Saha, Ole-Christoffer Granmo, Vladimir I. Zadorozhny, Morten Goodwin
TMs are a pattern recognition approach that uses finite state machines for learning and propositional logic to represent patterns.
6 code implementations • 7 Jan 2021 • Lei Jiao, Xuan Zhang, Ole-Christoffer Granmo, K. Darshana Abeyrathna
The Tsetlin Machine (TM) is a novel machine learning algorithm with several distinct properties, including transparent inference and learning using hardware-near building blocks.
5 code implementations • 17 Nov 2020 • Bimal Bhattarai, Ole-Christoffer Granmo, Lei Jiao
The mechanism uses the conjunctive clauses of the TM to measure to what degree a text matches the classes covered by the training data.
2 code implementations • 10 Sep 2020 • K. Darshana Abeyrathna, Bimal Bhattarai, Morten Goodwin, Saeed Gorji, Ole-Christoffer Granmo, Lei Jiao, Rupsa Saha, Rohan K. Yadav
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.
no code implementations • 28 Jul 2020 • Xuan Zhang, Lei Jiao, Ole-Christoffer Granmo, Morten Goodwin
The analysis of the convergence of the two basic operators lays the foundation for analyzing other logical operators.
no code implementations • 27 Jul 2020 • Christian D. Blakely, Ole-Christoffer Granmo
Additionally, the expressions capture the most important features of the model overall (global interpretability).
no code implementations • 4 Jul 2020 • K. Darshana Abeyrathna, Ole-Christoffer Granmo, Rishad Shafik, Alex Yakovlev, Adrian Wheeldon, Jie Lei, Morten Goodwin
However, TMs rely heavily on energy-costly random number generation to stochastically guide a team of Tsetlin Automata to a Nash Equilibrium of the TM game.
4 code implementations • 11 May 2020 • K. Darshana Abeyrathna, Ole-Christoffer Granmo, Morten Goodwin
Here, we address the accuracy-interpretability challenge in machine learning by equipping the TM clauses with integer weights.
1 code implementation • 7 Apr 2020 • Saeed Rahimi Gorji, Ole-Christoffer Granmo, Sondre Glimsdal, Jonathan Edwards, Morten Goodwin
Instead we use a simple look-up table that indexes the clauses on the features that falsify them.
no code implementations • 10 Feb 2020 • Rohan Kuamr Yadav, Lei Jiao, Ole-Christoffer Granmo, Morten Goodwin
Word Sense Disambiguation (WSD) is a longstanding unresolved task in Natural Language Processing.
4 code implementations • 4 Feb 2020 • K. Darshana Abeyrathna, Ole-Christoffer Granmo, Morten Goodwin
Although the RTM has solved non-linear regression problems with competitive accuracy, the resolution of the output is proportional to the number of clauses employed.
2 code implementations • 28 Nov 2019 • Adrian Phoulady, Ole-Christoffer Granmo, Saeed Rahimi Gorji, Hady Ahmady Phoulady
Finally, our novel sampling scheme reduced sample generation time by a factor of $7$.
Ranked #48 on Image Classification on MNIST
4 code implementations • 16 Sep 2019 • Saeed Rahimi Gorji, Ole-Christoffer Granmo, Adrian Phoulady, Morten Goodwin
The recently introduced Tsetlin Machine (TM) has provided competitive pattern recognition accuracy in several benchmarks, however, requires a 3-dimensional hyperparameter search.
no code implementations • 28 Aug 2019 • Jivitesh Sharma, Ole-Christoffer Granmo, Morten Goodwin
In this paper, we propose a model for the Environment Sound Classification Task (ESC) that consists of multiple feature channels given as input to a Deep Convolutional Neural Network (CNN) with Attention mechanism.
1 code implementation • 27 Jul 2019 • Per-Arne Andersen, Morten Goodwin, Ole-Christoffer Granmo
If these environment dynamics are adequately learned, a model-based approach is perhaps the most sample efficient method for learning agents to act in an environment optimally.
no code implementations • 15 Jul 2019 • Mehdi Ben Lazreg, Morten Goodwin, Ole-Christoffer Granmo
However, learning the graph structure is often complex, particularly when the graph is cyclic, and the transitions from one node to another are conditioned such as graphs used to represent a finite state machine.
8 code implementations • arXiv 2019 • Ole-Christoffer Granmo, Sondre Glimsdal, Lei Jiao, Morten Goodwin, Christian W. Omlin, Geir Thore Berge
Whereas the TM categorizes an image by employing each clause once to the whole image, the CTM uses each clause as a convolution filter.
Ranked #14 on Image Classification on Fashion-MNIST
1 code implementation • 10 May 2019 • K. Darshana Abeyrathna, Ole-Christoffer Granmo, Lei Jiao, Morten Goodwin
We achieve this by: (1) using the conjunctive clauses of the TM to capture arbitrarily complex patterns; (2) mapping these patterns to a continuous output through a novel voting and normalization mechanism; and (3) employing a feedback scheme that updates the TM clauses to minimize the regression error.
4 code implementations • 10 May 2019 • K. Darshana Abeyrathna, Ole-Christoffer Granmo, Xuan Zhang, Morten Goodwin
In this paper, we apply a new promising tool for pattern classification, namely, the Tsetlin Machine (TM), to the field of disease forecasting.
1 code implementation • 2 Oct 2018 • Per-Arne Andersen, Morten Goodwin, Ole-Christoffer Granmo
It is likely that the solution to these problems lies in short- and long-term planning, exploration and memory management for reinforcement learning algorithms.
1 code implementation • 12 Sep 2018 • Geir Thore Berge, Ole-Christoffer Granmo, Tor Oddbjørn Tveit, Morten Goodwin, Lei Jiao, Bernt Viggo Matheussen
The Tsetlin Machine either performs on par with or outperforms all of the evaluated methods on both the 20 Newsgroups and IMDb datasets, as well as on a non-public clinical dataset.
1 code implementation • 15 Aug 2018 • Per-Arne Andersen, Morten Goodwin, Ole-Christoffer Granmo
Reinforcement learning (RL) is an area of research that has blossomed tremendously in recent years and has shown remarkable potential for artificial intelligence based opponents in computer games.
15 code implementations • 4 Apr 2018 • Ole-Christoffer Granmo
Our theoretical analysis establishes that the Nash equilibria of the game align with the propositional formulas that provide optimal pattern recognition accuracy.
Ranked #49 on Image Classification on MNIST
no code implementations • 26 Jan 2018 • Per-Arne Andersen, Morten Goodwin, Ole-Christoffer Granmo
This paper introduces the Flash Reinforcement Learning platform (FlashRL) which attempts to fill this gap by providing an environment for thousands of Flash games on a novel platform for Flash automation.
no code implementations • 17 Dec 2017 • Per-Arne Andersen, Morten Goodwin, Ole-Christoffer Granmo
We propose a game environment in between Atari 2600 and Starcraft II, particularly targeting Deep Reinforcement Learning algorithm research.
no code implementations • 5 Aug 2017 • Sondre Glimsdal, Ole-Christoffer Granmo
In this paper, we address a particularly intriguing variant of the multi-armed bandit problem, referred to as the {\it Stochastic Point Location (SPL) Problem}.
no code implementations • 11 Jul 2017 • Sondre Glimsdal, Ole-Christoffer Granmo
This problem is generally referred to as the Object Partitioning Problem (OPP) and is known to be NP-hard.
no code implementations • 28 May 2017 • Ole-Christoffer Granmo
Due to the pervasiveness of bandit based optimisation, our scheme opens up for improved performance both in meta-optimisation and in applications where gradient related information is readily available.