Search

The Tsetlin Machine (TM) is a recent machine learning algorithm with several distinct properties, such as interpretability, simplicity, and hardware-friendliness. Although numerous empirical evaluations report on its performance, the mathematical analysis of its convergence is still open. In this article, we analyze the convergence of the TM with only one clause involved for classification. More specifically, we examine two basic logical operators, namely, the "IDENTITY"- and "NOT" operators. Our analysis reveals that the TM, with just one clause, can converge correctly to the intended logical operator, learning from training data over an infinite time horizon. Besides, it can capture arbitrarily rare patterns and select the most accurate one when two candidate patterns are incompatible, by configuring a granularity parameter. The analysis of the convergence of the two basic operators lays the foundation for analyzing other logical operators. These analyses altogether, from a mathematical perspective, provide new insights on why TMs have obtained state-of-the-art performance on several pattern recognition problems., Comment: 23 pages, 10 figures

We focus on the important problem of emergency evacuation, which clearly could benefit from reinforcement learning that has been largely unaddressed. Emergency evacuation is a complex task which is difficult to solve with reinforcement learning, since an emergency situation is highly dynamic, with a lot of changing variables and complex constraints that makes it difficult to train on. In this paper, we propose the first fire evacuation environment to train reinforcement learning agents for evacuation planning. The environment is modelled as a graph capturing the building structure. It consists of realistic features like fire spread, uncertainty and bottlenecks. We have implemented the environment in the OpenAI gym format, to facilitate future research. We also propose a new reinforcement learning approach that entails pretraining the network weights of a DQN based agents to incorporate information on the shortest path to the exit. We achieved this by using tabular Q-learning to learn the shortest path on the building model's graph. This information is transferred to the network by deliberately overfitting it on the Q-matrix. Then, the pretrained DQN model is trained on the fire evacuation environment to generate the optimal evacuation path under time varying conditions. We perform comparisons of the proposed approach with state-of-the-art reinforcement learning algorithms like PPO, VPG, SARSA, A2C and ACKTR. The results show that our method is able to outperform state-of-the-art models by a huge margin including the original DQN based models. Finally, we test our model on a large and complex real building consisting of 91 rooms, with the possibility to move to any other room, hence giving 8281 actions. We use an attention based mechanism to deal with large action spaces. Our model achieves near optimal performance on the real world emergency environment., 21 pages, 14 figures, 4 tables

MedAI: Transparency in Medical Image Segmentation is a challenge held for the first time at the Nordic AI Meet that focuses on medical image segmentation and transparency in machine learning (ML)-based systems. We propose three tasks to meet specific gastrointestinal image segmentation challenges collected from experts within the field, including two separate segmentation scenarios and one scenario on transparent ML systems. The latter emphasizes the need for explainable and interpretable ML algorithms. We provide a development dataset for the participants to train their ML models, tested on a concealed test dataset.

Reinforcement learning has shown to be profoundly successful at learning optimal policies for simulated environments using distributed training with extensive compute capacity. Model-free reinforcement learning uses the notion of trial and error, where the error is a vital part of learning the agent to behave optimally. In mission-critical, real-world environments, there is little tolerance for failure and can cause damaging effects on humans and equipment. In these environments, current state-of-the-art reinforcement learning approaches are not sufficient to learn optimal control policies safely. On the other hand, model-based reinforcement learning tries to encode environment transition dynamics into a predictive model. The transition dynamics describes the mapping from one state to another, conditioned on an action. If this model is accurate enough, the predictive model is sufficient to train agents for optimal behavior in real environments. This paper presents the Dreaming Variational Autoencoder (DVAE) for safely learning good policies with a significantly lower risk of catastrophes occurring during training. The algorithm combines variational autoencoders, risk-directed exploration, and curiosity to train deep-q networks inside ”dream” states. We introduce a novel environment, ASRS-Lab, for research in the safe learning of autonomous vehicles in grid-based warehousing. The work shows that the proposed algorithm has better sample efficiency with similar performance to novel model-free deep reinforcement learning algorithms while maintaining safety during training.

Physical inactivity increases the risk of many adverse health conditions, including the world’s major non-communicable diseases, such as coronary heart disease, type 2 diabetes, and breast and colon cancers, shortening life expectancy. There are minimal medical care and personal trainers’ methods to monitor a patient’s actual physical activity types. To improve activity monitoring, we propose an artificial-intelligence-based approach to classify the physical movement activity patterns. In more detail, we employ two deep learning (DL) methods, namely a deep feed-forward neural network (DNN) and a deep recurrent neural network (RNN) for this purpose. We evaluate the proposed models on two physical movement datasets collected from several volunteers who carried tri-axial accelerometer sensors. The first dataset is from the UCI machine learning repository, which contains 14 different activities-of-daily-life (ADL) and is collected from 16 volunteers who carried a single wrist-worn tri-axial accelerometer. The second dataset includes ten other ADLs and is gathered from 8 volunteers who placed the sensors on their hips. Our experiment results show that the RNN model provides the accuracy performance compared to the state-of-the-art methods in classifying the fundamental movement patterns with an overall accuracy of 84.89% and an overall F1-score of 82.56%. Our results indicate that the proposed method will provide the medical doctors and trainers a promising way to precisely track and understand a patient’s physical activities for better treatment.

Polysomnography (PSG), the gold standard for sleep stage classification, requires a sleep expert for scoring and is both resource-intensive and expensive. Many researchers currently focus on the real-time classification of the sleep stages based on biomedical signals, such as Electroencephalograph (EEG) and electrooculography (EOG). However, most of the research work is based on machine learning models with multiple signal inputs or hand-engineered features requiring prior knowledge of the sleep domain. We propose a novel encoded Time-Distributed Convolutional Neural Network (TDConvNet) to automatically classify sleep stages based on a single raw PSG signal. The TDConvNet can infer sleep stages in just 30-second epochs for the 5-stage sleep classification using a single EEG or EOG signal from the open Sleep-EDF dataset. We evaluated the TDConvNet performance on EEGs and EOG signals. The evaluation results show that the proposed method achieved the best performance with the EEG Fpz-Cz signal (0.85) compared to current literature, followed by EEG Pz-Oz (0.84) and EOG horizontal (0.82). The source code is available at https://github.com/michealdutt/TDConvNet.

The Convolutional Tsetlin Machine (CTM), a variant of Tsetlin Machine (TM), represents patterns as straightforward AND-rules, to address the high computational complexity and the lack of interpretability of Convolutional Neural Networks (CNNs). CTM has shown competitive performance on MNIST, Fashion-MNIST, and Kuzushiji-MNIST pattern classification benchmarks, both in terms of accuracy and memory footprint. In this paper, we propose the Convolutional Regression Tsetlin Machine (C-RTM) that extends the CTM to support continuous output problems in image analysis. C-RTM identifies patterns in images using the convolution operation as in the CTM and then maps the identified patterns into a real-valued output as in the Regression Tsetlin Machine (RTM). The C-RTM thus unifies the two approaches. We evaluated the performance of C-RTM using 72 different artificial datasets, with and without noise in the training data. Our empirical results show the competitive performance of C-RTM compared to two standard CNNs. Additionally, the interpretability of the identified sub-patterns by C-RTM clauses is analyzed and discussed.

In this paper, we introduce a novel application of using scene semantic image segmentation for fire emergency situation analysis. To analyse a fire emergency scene, we propose to use deep convolutional image segmentation networks to identify and classify objects in a scene based on their build material and their vulnerability to catch fire. We introduce our own fire emergency scene segmentation dataset for this purpose. It consists of real world images with objects annotated on the basis of their build material. We use state-of-the-art segmentation models: DeepLabv3, DeepLabv3+, PSPNet, FCN, SegNet and UNet to compare and evaluate their performance on the fire emergency scene parsing task. During inference time, we only run the encoder (backbone) network to determine whether there is a fire or not in the image. If there is a fire, only then the decoder is activated to segment the emergency scene. This results in dispensing with unnecessary computation, i.e. the decoder. We achieve this by using multitask learning. We show the importance of transfer learning and the difference in performance between models pretrained on different benchmark datasets. The results show that segmentation models can accurately analyse an emergency situation, if properly trained to do so. Our fire emergency scene parsing dataset is available here: https://github.com/cair.

Whole gland (WG) segmentation of the prostate plays a crucial role in detection, staging and treatment planning of prostate cancer (PCa). Despite promise shown by deep learning (DL) methods, they rely on the availability of a considerable amount of annotated data. Augmentation techniques such as translation and rotation of images present an alternative to increase data availability. Nevertheless, the amount of information provided by the transformed data is limited due to the correlation between the generated data and the original. Based on the recent success of generative adversarial networks (GAN) in producing synthetic images for other domains as well as in the medical domain, we present a pipeline to generate WG segmentation masks and synthesize T2-weighted MRI of the prostate based on a publicly available multi-center dataset. Following, we use the generated data as a form of data augmentation. Results show an improvement in the quality of the WG segmentation when compared to standard augmentation techniques., Comment: 5 pages. Accepted as a full paper at the International Symposium on Biomedical Imaging (ISBI) 2021

Deep Reinforcement Learning (RL) is unquestionably a robust framework to train autonomous agents in a wide variety of disciplines. However, traditional deep and shallow model-free RL algorithms suffer from low sample efficiency and inadequate generalization for sparse state spaces. The options framework with temporal abstractions [18] is perhaps the most promising method to solve these problems, but it still has noticeable shortcomings. It only guarantees local convergence, and it is challenging to automate initiation and termination conditions, which in practice are commonly hand-crafted.

Aspect-based sentiment analysis (ABSA) aims at identifying fine-grained polarity of opinion associated with a given aspect word. Several existing articles demonstrated promising ABSA accuracy using positional embedding to show the relationship between an aspect word and its context. In most cases, the positional embedding depends on the distance between the aspect word and the remaining words in the context, known as the position index sequence. However, these techniques usually employ both complex preprocessing approaches with additional trainable positional embedding and complex architectures to obtain the state-of-the-art performance. In this paper, we simplify preprocessing by including polarity lexicon replacement and masking techniques that carry the information of the aspect word’s position and eliminate the positional embedding. We then adopt a novel and concise architecture using two Bidirectional GRU along with an attention layer to classify the aspect based on its context words. Experiment results show that the simplified preprocessing and the concise architecture significantly improve the accuracy of the publicly available ABSA datasets, obtaining 81.37%, 75.39%, 80.88%, and 89.30% in restaurant 14, laptop 14, restaurant 15, and restaurant 16 respectively.

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.

The Tsetlin Machine (TM) is a recent interpretable machine learning algorithm that requires relatively modest computational power, yet attains competitive accuracy in several benchmarks. TMs are inherently binary; however, many machine learning problems are continuous. While binarization of continuous data through brute-force thresholding has yielded promising accuracy, such an approach is computationally expensive and hinders extrapolation. In this paper, we address these limitations by standardizing features to support scale shifts in the transition from training data to real-world operation, typical for e.g. forecasting. For scalability, we employ sampling to reduce the number of binarization thresholds, relying on stratification to minimize loss of accuracy. We evaluate the approach empirically using two artificial datasets before we apply the resulting TM to forecast dengue outbreaks in the Philippines using the spatiotemporal properties of the data. Our results show that the loss of accuracy due to threshold sampling is insignificant. Furthermore, the dengue outbreak forecasts made by the TM are more accurate than those obtained by Support Vector Machines (SVMs), Decision Trees (DTs), and several multi-layered Artificial Neural Networks (ANNs), both in terms of forecasting precision and Fl-score.

Tsetlin machines (TMs) are a promising approach to machine learning that uses Tsetlin Automata to produce patterns in propositional logic, leading to binary (hard) classifications. In many applications, however, one needs to know the confidence of classifications, e.g. to facilitate risk management. In this paper, we propose a novel scheme for measuring TM confidence based on the logistic function, calculated from the propositional logic patterns that match the input. We then use this scheme to trade off precision against recall, producing area under receiver operating characteristic curves (AUC) for TMs. Empirically, using four real-world datasets, we show that AUC is a more sensitive measure of TM performance compared to Accuracy. Further, the AUC-based evaluations show that the TM performs on par or better than widely used machine learning algorithms. We thus believe our scheme will make the TM more suitable for use in decision support, where the user needs to inspect and validate predictions, in particular, those being uncertain.

In this paper, we propose a generic emergency detection system using only the sound produced in the environment. For this task, we employ multiple audio feature extraction techniques like the mel-frequency cepstral coefficients, gammatone frequency cepstral coefficients, constant Q-transform and chromagram. After feature extraction, a deep convolutional neural network (CNN) is used to classify an audio signal as a potential emergency situation or not. The entire model is based on our previous work that sets the new state of the art in the environment sound classification (ESC) task (Our paper is under review in the IEEE/ACM Transactions on Audio, Speech and Language Processing and also available here https://arxiv.org/abs/1908.11219.) We combine the benchmark ESC datasets: UrbanSound8K and ESC-50 (ESC-10 is a subset of ESC-50) and reduce the problem to a binary classification problem. This is done by aggregating sound classes such as sirens, fire crackling, glass breaking, gunshot as the emergency class and others as normal. Even though there are only two classes to distinguish, they are highly imbalanced. To overcome this difficulty, we introduce class weights in calculating the loss while training the model. Our model is able to achieve \(99.56\%\) emergency detection accuracy.

Deep reinforcement learning is considered promising for many energy cost optimization tasks in smart buildings. How-ever, agent learning, in this context, is sometimes unstable and unpredictable, especially when the environments are complex. In this paper, we examine deep Reinforcement Learning (RL) algorithms developed for game play applied to a battery control task with an energy cost optimization objective. We explore how agent behavior and hyperparameters can be analyzed in a simplified environment with the goal of modifying the algorithms for increased stability. Our modified Deep Deterministic Policy Gradient (DDPG) agent is able to perform consistently close to the optimum over multiple training sessions with a maximum cost reduction of 25 % and an average cost reduction of 99 % of the maximum in the simplified environment. DDPG is an actor-critic RL algorithm consisting of four neural networks - the actor and critic, main and target, networks. When environment complexity is increased, the DDPG agent performance decreases and a modified Twin Delayed DDPG (TD3) agent is utilized to achieve an average of 99.9 % of the optimal result. The TD3 algorithm uses two main critic networks to avoid known value overestimation bias.

Despite significant effort, building models that are both interpretable and accurate is an unresolved challenge for many pattern recognition problems. In general, rule-based and linear models lack accuracy, while deep learning interpretability is based on rough approximations of the underlying inference. Using a linear combination of conjunctive clauses in propositional logic, Tsetlin Machines (TMs) have shown competitive performance on diverse benchmarks. However, to do so, many clauses are needed, which impacts interpretability. Here, we address the accuracy-interpretability challenge in machine learning by equipping the TM clauses with integer weights. The resulting Integer Weighted TM (IWTM) deals with the problem of learning which clauses are inaccurate and thus must team up to obtain high accuracy as a team (low weight clauses), and which clauses are sufficiently accurate to operate more independently (high weight clauses). Since each TM clause is formed adaptively by a team of Tsetlin Automata, identifying effective weights becomes a challenging online learning problem. We address this problem by extending each team of Tsetlin Automata with a stochastic searching on the line (SSL) automaton. In our novel scheme, the SSL automaton learns the weight of its clause in interaction with the corresponding Tsetlin Automata team, which, in turn, adapts the composition of the clause by the adjusting weight. We evaluate IWTM empirically using five datasets, including a study of interpetability. On average, IWTM uses 6.5 times fewer literals than the vanilla TM and 120 times fewer literals than a TM with real-valued weights. Furthermore, in terms of average F1-Score, IWTM outperforms simple Multi-Layered Artificial Neural Networks, Decision Trees, Support Vector Machines, K-Nearest Neighbor, Random Forest, XGBoost, Explainable Boosting Machines, and standard and real-value weighted TMs., 20 pages, 10 figures

Human mistakes in traffic often have terrible consequences. The long-awaited introduction of self-driving vehicles may solve many of the problems with traffic, but much research is still needed before cars are fully autonomous.In this paper, we propose a new Road Detection algorithm using online supervised learning based on a Neural Network architecture. This algorithm is designed to support a Reinforcement Learning algorithm (for example, the standard Proximal Policy Optimization or PPO) by detecting when the car is in an adverse condition. Specifically, the PPO gets a penalty whenever the virtual automobile gets stuck or drives off the road with any of its four wheels.Initial experiments show significantly improved results for PPO when using our Road Detection algorithm, as compared to not using any form of Road Detection.In fact, without this detection algorithm, the vehicle often gets into non-terminating loops (for example, driving into the dividers, getting stuck, or driving into a pit).

Tsetlin Machines (TMs) are an interpretable pattern recognition approach that captures patterns with high discriminative power from data. Patterns are represented as conjunctive clauses in propositional logic, produced using bandit-learning in the form of Tsetlin Automata. In this work, we propose a TM-based approach to two common Natural Language Processing (NLP) tasks, viz. Sentiment Analysis and Semantic Relation Categorization. By performing frequent itemset mining on the patterns produced, we show that they follow existing expert-verified rule-sets or lexicons. Further, our comparison with other widely used machine learning techniques indicates that the TM approach helps maintain interpretability without compromising accuracy – a result we believe has far-reaching implications not only for interpretable NLP but also for interpretable AI in general.

Reinforcement Learning (RL) is a general framework concerned with an agent that seeks to maximize rewards in an environment. The learning typically happens through trial and error using explorative methods, such as \(\epsilon \)-greedy. There are two approaches, model-based and model-free reinforcement learning, that show concrete results in several disciplines. Model-based RL learns a model of the environment for learning the policy while model-free approaches are fully explorative and exploitative without considering the underlying environment dynamics. Model-free RL works conceptually well in simulated environments, and empirical evidence suggests that trial and error lead to a near-optimal behavior with enough training. On the other hand, model-based RL aims to be sample efficient, and studies show that it requires far less training in the real environment for learning a good policy.

Author's accepted manuscript Industrial cooling systems consume large quantities of energy with highly variable power demand. To reduce environmental impact and overall energy consumption, and to stabilize the power requirements, it is recommended to recover surplus heat, store energy, and integrate renewable energy production. To control these operations continuously in a complex energy system, an intelligent energy management system can be employed using operational data and machine learning. In this work, we have developed an artificial neural network based technique for modelling operational CO2 refrigerant based industrial cooling systems for embedding in an overall energy management system. The operating temperature and pressure measurements, as well as the operating frequency of compressors, are used in developing operational model of the cooling system, which outputs electrical consumption and refrigerant mass flow without the need for additional physical measurements. The presented model is superior to a generalized theoretical model, as it learns from data that includes individual compressor type characteristics. The results show that the presented approach is relatively precise with a Mean Average Percentage Error (MAPE) as low as 5%, using low resolution and asynchronous data from a case study system. The developed model is also tested in a laboratory setting, where MAPE is shown to be as low as 1.8%.

Mitigating misinformation on social media is an unresolved challenge, particularly because of the complexity of information dissemination. To this end, Multivariate Hawkes Processes (MHP) have become a fundamental tool because they model social network dynamics, which facilitates execution and evaluation of mitigation policies. In this paper, we propose a novel light-weight intervention-based misinformation mitigation framework using decentralized Learning Automata (LA) to control the MHP. Each automaton is associated with a single user and learns to what degree that user should be involved in the mitigation strategy by interacting with a corresponding MHP, and performing a joint random walk over the state space. We use three Twitter datasets to evaluate our approach, one of them being a new COVID-19 dataset provided in this paper. Our approach shows fast convergence and increased valid information exposure. These results persisted independently of network structure, including networks with central nodes, where the latter could be the root of misinformation. Further, the LA obtained these results in a decentralized manner, facilitating distributed deployment in real-life scenarios.

Due to the high energy consumption and scalability challenges of deep learning, there is a critical need to shift research focus towards dealing with energy consumption constraints. Tsetlin Machines (TMs) are a recent approach to machine learning that has demonstrated significantly reduced energy usage compared to neural networks alike, while performing competitively accuracy-wise on several benchmarks. However, TMs rely heavily on energy-costly random number generation to stochastically guide a team of Tsetlin Automata (TA) to a Nash Equilibrium of the TM game. In this paper, we propose a novel finite-state learning automaton that can replace the TA in TM learning, for increased determinism. The new automaton uses multi-step deterministic state jumps to reinforce sub-patterns. Simultaneously, flipping a coin to skip every d’th state update ensures diversification by randomization. The d-parameter thus allows the degree of randomization to be finely controlled. E.g., \(d=1\) makes every update random and \(d=\infty \) makes the automaton completely deterministic. Our empirical results show that, overall, only substantial degrees of determinism reduces accuracy. Energy-wise, random number generation constitutes switching energy consumption of the TM, saving up to 11 mW power for larger datasets with high d values. We can thus use the new d-parameter to trade off accuracy against energy consumption, to facilitate low-energy machine learning.

A wide range of applications in marine ecology extensively uses underwater cameras. Still, to efficiently process the vast amount of data generated, we need to develop tools that can automatically detect and recognize species captured on film. Classifying fish species from videos and images in natural environments can be challenging because of noise and variation in illumination and the surrounding habitat. In this paper, we propose a two-step deep learning approach for the detection and classification of temperate fishes without pre-filtering. The first step is to detect each single fish in an image, independent of species and sex. For this purpose, we employ the You Only Look Once (YOLO) object detection technique. In the second step, we adopt a Convolutional Neural Network (CNN) with the Squeeze-and-Excitation (SE) architecture for classifying each fish in the image without pre-filtering. We apply transfer learning to overcome the limited training samples of temperate fishes and to improve the accuracy of the classification. This is done by training the object detection model with ImageNet and the fish classifier via a public dataset (Fish4Knowledge), whereupon both the object detection and classifier are updated with temperate fishes of interest. The weights obtained from pre-training are applied to post-training as a priori. Our solution achieves the state-of-the-art accuracy of 99.27\% on the pre-training. The percentage values for accuracy on the post-training are good; 83.68\% and 87.74\% with and without image augmentation, respectively, indicating that the solution is viable with a more extensive dataset., Comment: arXiv admin note: substantial text overlap with arXiv:1904.02768

One of the most well-known and complex applications of artificial intelligence (AI) is Alzheimer’s detection, which lies in the field of medical imaging. The complexity in this task lies in the three-dimensional structure of the MRI scan images. In this paper, we propose to use 3D Convolutional Neural Networks (3D-CNN) for Alzheimer’s detection. 3D-CNNs have been a popular choice for this task. The novelty in our paper lies in the fact that we use a deeper 3D-CNN consisting of 10 layers. Also, with effectively training our model consisting of Batch Normalization layers that provide a regularizing effect, we don’t have to use any transfer learning. We also use the simple data augmentation technique of flipping. Our model is trained for binary classification that distinguishes between Alzheimer’s and normal, as well as multiclass classification consisting of Alzheimer’s, Mild Cognitive Impairment, and normal classes. We tested our model on the ADNI dataset and achieved 94.17% and 89.14% accuracy for binary classification and multiclass classification, respectively.

Whenever a disaster occurs, users in social media, sensors, cameras, satellites, and the like generate vast amounts of data. Emergency responders and victims use this data for situational awareness, decision-making, and safe evacuations. However, making sense of the generated information under time-bound situations is a challenging task as the amount of data can be significant, and there is a need for intelligent systems to analyze, process, and visualize it. With recent advancements in Artificial Intelligence (AI), numerous researchers have begun exploring AI, machine learning (ML), and deep learning (DL) techniques for big data analytics in managing disasters efficiently. This paper adopts a systematic literature approach to report on the application of AI, ML, and DL in disaster management. Through a systematic review process, we identified one relevant hundred publications. After that, we analyzed all the identified papers and concluded that most of the reviewed articles used AI, ML, and DL methods on social media data, satellite data, sensor data, and historical data for classification and prediction. The most common algorithms are support vector machines (SVM), Naive Bayes (NB), Random Forest (RF), Convolutional Neural Networks (CNN), Artificial neural networks (ANN), Natural language processing techniques (NLP), Latent Dirichlet Allocation (LDA), K-nearest neighbor (KNN), and Logistic Regression (LR).

During to past few years, social media have gained a pivotal role in crisis communication. Its usage has ranged from informing the public about the status of a crisis and what precaution need to be taken, to family members checking on the safety of loved ones. Despite the widespread use of social media in crises situations and the clear potential benefit from collecting potential critical information from social media, emergency management services (EMSs) are still reluctant to use social media as a source of information to improve their situational awareness. One of the reasons for the reluctance is that crises management are typically overloaded with information. Adding social media will only increase the information overload, and the EMSs risk being provided with more and possibly irrelevant information from social media sources. Furthermore, most automated social media analysis platforms are designed exclusively to classifying messages into crisis and not crisis-related categories. The platforms do not take into account the degree of relevance of social media information to the EMSs. Such relevance further depends on the crisis' status at a certain point in time, and the information gathered up to that point. This paper proposes an intelligent information retrieval framework from social media during crises. The developed framework combines two components: The first component classifies social media messages into separate topics representing the information needed in different situations. The second component decides which information to retrieve based on the information available and the status of the crisis. The framework is evaluated through a survey of EMSs. They agree with the framework in 70.85% of the cases about what information is needed.

Recent advances in intrusion detection systems based on machine learning have indeed outperformed other techniques, but struggle with detecting multiple classes of attacks with high accuracy. We propose a method that works in three stages. First, the ExtraTrees classifier is used to select relevant features for each type of attack individually for each (ELM). Then, an ensemble of ELMs is used to detect each type of attack separately. Finally, the results of all ELMs are combined using a softmax layer to refine the results and increase the accuracy further. The intuition behind our system is that multi-class classification is quite difficult compared to binary classification. So, we divide the multi-class problem into multiple binary classifications. We test our method on the UNSW and KDDcup99 datasets. The results clearly show that our proposed method is able to outperform all the other methods, with a high margin. Our system is able to achieve 98.24% and 99.76% accuracy for multi-class classification on the UNSW and KDDcup99 datasets, respectively. Additionally, we use the weighted extreme learning machine to alleviate the problem of imbalance in classification of attacks, which further boosts performance. Lastly, we implement the ensemble of ELMs in parallel using GPUs to perform intrusion detection in real time.

Humanitarian aid workers who try to provide aid to the most vulnerable populations in the Middle East or Africa are risking their own lives and safety to help others. The current lack of a collaborative real-time information system to predict threats prevents responders and local partners from developing a shared understanding of potentially threatening situations, causing increased response times and leading to inadequate protection. To solve this problem, this paper presents a threat detection and decision support system that combines knowledge and information from a network of responders with automated and modular threat detection. The system consists of three parts. It first collects textual information, ranging from social media, and online news reports to reports and text messages from a decentralized network of humanitarian staff. Second, the system uses deep neural network techniques to automatically detects a threat or incident and provide information including location, threat category, and casualties. Third, given the type of threat and the information extracted by the NER, a feedforward network proposes a mitigation plan based on humanitarian standard operating procedures. The classified information is rapidly redistributed to potentially affected humanitarian workers at any level. The system testing results show a high precision of 0.91 and 0.98 as well as an F-measure of 0.87 and 0.88 in detecting the threats and decision support respectively. We thus combine the collaborative intelligence of a decentralized network of aid workers with the power of deep neural networks. CCS CONCEPTS • Computing methodologies → Neural networks.

Most of the Nordic holiday resorts are in rural area with low capacity distributed network. The rural area network is weak and needs capacity expansion planning as the load demand of this area are going to increase due to penetration of electric vehicles and heat pumps. Such type of rural network can also be operated as a micro-grid, and therefore load analysis is required for appropriate operation. The load analysis will also be useful for finding proper sizing of distributed energy resources including energy storage. In this work, load demand analysis of a typical Nordic holiday resorts, connected in rural grid, is presented to find out the load variation during the usage periods. The load analysis is targeted for demand prediction. The demand forecasting has been considered through integrating Regression Tools with Artificial Neural Networks due to the low amount of data available from the Holiday Resorts. Collected data is from a rural area in Norway consisting of 125 holiday cabins, with maximum load of 478 kW in the period of 2014 to 2018. This work is presenting the analysis on the total electric load consumption of cabins during typical short and long term holidays. It is observed, during the longer time holiday period, the loads are significantly higher compared to shorter time holiday period. Prediction analysis shows that the MAPE is relatively higher compare to predicted results in higher load area. Through analysis, it is observed that the curvature of the maximum peak demand is unfitting the predictive outcome. To overcome this problem the finite gradient by autoregression, has been used in this work.

Teeth are some of the most resilient tissues of the human body. Because of their placement, teeth often yield intact indicators even when other metrics, such as finger prints and DNA, are missing. Forensics on dental identification is now mostly manual work which is time and resource intensive. Systems for automated human identification from dental X-ray images have the potential to greatly reduce the necessary efforts spent on dental identification, but it requires a system with high stability and accuracy so that the results can be trusted. This paper proposes a new system for automated dental X-ray identification. The scheme extracts tooth and dental work contours from the X-ray images and uses the Hausdorff-distance measure for ranking persons. This combination of state-of-the-art approaches with a novel lowest cost path-based method for separating a dental X-ray image into individual teeth, is able to achieve comparable and better results than what is available in the literature. The proposed scheme is fully functional and is used to accurately identify people within a real dental database. The system is able to perfectly separate 88.7% of the teeth in the test set. Further, in the verification process, the system ranks the correct person in top in 86% of the cases, and among the top five in an astonishing 94% of the cases. The approach has compelling potential to significantly reduce the time spent on dental identification.

Predicting and mitigating demand peaks in electrical networks has become a prevalent research topic. Demand peaks pose a particular challenge to energy companies because these are difficult to foresee and require the net to support abnormally high consumption levels. In smart energy grids, time-differentiated pricing policies that increase the energy cost for the consumers during peak periods, and load balancing are examples of simple techniques for peak regulation. In this paper, we tackle the task of predicting power peaks prior to their actual occurrence in the context of a pilot Norwegian smart grid network.

Medical applications challenge today's text categorization techniques by demanding both high accuracy and ease-of-interpretation. Although deep learning has provided a leap ahead in accuracy, this leap comes at the sacrifice of interpretability. To address this accuracy-interpretability challenge, we here introduce, for the first time, a text categorization approach that leverages the recently introduced Tsetlin Machine. In all brevity, we represent the terms of a text as propositional variables. From these, we capture categories using simple propositional formulae, such as: if "rash" and "reaction" and "penicillin" then Allergy. The Tsetlin Machine learns these formulae from a labelled text, utilizing conjunctive clauses to represent the particular facets of each category. Indeed, even the absence of terms (negated features) can be used for categorization purposes. Our empirical comparison with Na\"ive Bayes, decision trees, linear support vector machines (SVMs), random forest, long short-term memory (LSTM) neural networks, and other techniques, is quite conclusive. 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. On average, the Tsetlin Machine delivers the best recall and precision scores across the datasets. Finally, our GPU implementation of the Tsetlin Machine executes 5 to 15 times faster than the CPU implementation, depending on the dataset. We thus believe that our novel approach can have a significant impact on a wide range of text analysis applications, forming a promising starting point for deeper natural language understanding with the Tsetlin Machine., Comment: 10 pages, 4 figures

The recently introduced Tsetlin Machine (TM) has provided competitive pattern recognition accuracy in several benchmarks, however, requires a 3-dimensional hyperparameter search. In this paper, we introduce the Multigranular Tsetlin Machine (MTM). The MTM eliminates the specificity hyperparameter, used by the TM to control the granularity of the conjunctive clauses that it produces for recognizing patterns. Instead of using a fixed global specificity, we encode varying specificity as part of the clauses, rendering the clauses multigranular. This makes it easier to configure the TM because the dimensionality of the hyperparameter search space is reduced to only two dimensions. Indeed, it turns out that there is significantly less hyper-parameter tuning involved in applying the MTM to new problems. Further, we demonstrate empirically that the MTM provides similar performance to what is achieved with a finely specificity-optimized TM, by comparing their performance on both synthetic and real-world datasets.

Large scale cooling installations usually have high energy consumption and fluctuating power demands. There are several ways to control energy consumption and power requirements through intelligent energy and power management, such as utilizing excess heat, thermal energy storage and local renewable energy sources. Intelligent energy and power management in an operational setting is only possible if the time-varying performance of the individual components of the energy system is known. This paper presents an approach to model the compressors in an industrial, operational two-stage cooling system, with CO\(_2\) as the working fluid, located in an advanced food distribution warehouse in Norway. An artificial neural network is adopted to model the compressors using the operational data. The model is trained with cooling medium evaporation and condensation temperature, suction gas temperature and compressor operating frequency, and outputs electrical power load and cooling load. The best results are found by using a single hidden layer with 45 hidden neurons and a hyperbolic tangent activation function trained with the Adam optimizer, with a resulting mean squared error as low as 0.08% for both the training and validation data sets. The trained model will be part of a system implemented in a real-world setting to determine the cooling load, compressor power load, and coefficient of performance. An intelligent energy management system will utilize the model for energy and power optimization of the cooling system by storing energy in a thermal energy storage, using predictions of energy demand and cooling system performance.

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. The novelty of the paper lies in using multiple feature channels consisting of Mel-Frequency Cepstral Coefficients (MFCC), Gammatone Frequency Cepstral Coefficients (GFCC), the Constant Q-transform (CQT) and Chromagram. Such multiple features have never been used before for signal or audio processing. And, we employ a deeper CNN (DCNN) compared to previous models, consisting of spatially separable convolutions working on time and feature domain separately. Alongside, we use attention modules that perform channel and spatial attention together. We use some data augmentation techniques to further boost performance. Our model is able to achieve state-of-the-art performance on all three benchmark environment sound classification datasets, i.e. the UrbanSound8K (97.52%), ESC-10 (95.75%) and ESC-50 (88.50%). To the best of our knowledge, this is the first time that a single environment sound classification model is able to achieve state-of-the-art results on all three datasets. For ESC-10 and ESC-50 datasets, the accuracy achieved by the proposed model is beyond human accuracy of 95.7% and 81.3% respectively., Comment: Re-checking results

Deep reinforcement learning has over the past few years shown great potential in learning near-optimal control in complex simulated environments with little visible information. Rainbow (Q-Learning) and PPO (Policy Optimisation) have shown outstanding performance in a variety of tasks, including Atari 2600, MuJoCo, and Roboschool test suite. Although these algorithms are fundamentally different, both suffer from high variance, low sample efficiency, and hyperparameter sensitivity that, in practice, make these algorithms a no-go for critical operations in the industry.

Measuring similarities between strings is central for many established and fast-growing research areas, including information retrieval, biology, and natural-language processing. The traditional approach to string similarity measurements is to define a metric with respect to a word space that quantifies and sums up the differences between characters in two strings; surprisingly, these metrics have not evolved a great deal over the past few decades. Indeed, the majority of them are still based on making a simple comparison between character and character distributions without considering the words context. This paper proposes a string metric that encompasses similarities between strings based on (1) the character similarities between the words, including non-standard and standard spellings of the same words, and (2) the context of these words. We propose a neural network composed of a denoising autoencoder and what we call a context encoder, both specifically designed to find similarities between the words based on their context. Experimental results show that the resulting metrics have succeeded in 85.4% of the cases in finding the correct version of a non-standard spelling among the closest words, compared to 63.2% using the established Normalised-Levenshtein distance. We also show that by employing our approach, the words used in similar context are calculated to be more similar than words with different contexts, which is a desirable property lacking in established string metrics.

When any kind of disaster occurs, victims who are directly and indirectly affected by the disaster often post vast amount of data (e.g., images, text, speech, video) using numerous social media platforms. This is because social media has recently become a primary communication channel among people to report either to public or to emergency responders (ERs). ERs, who are from various emergency response organizations (EROs), usually consider to gain awareness of the situation in order to respond to occurred disaster. However, with the occurrence of the disaster, within minutes, the social media platforms are flooded with various kinds of data which become overwhelmed for ERs with big data. Further, in this posted data, there may be majority of the data consist of redundant and irrelevant content. With this, it becomes challenging for ERs to make sense and take decisions of/on the available big data. Despite recent advances in the technology, processing and analyzing of the disaster related social media big data remains a challenging task. Hence, in this paper, we focus on presenting an initial analysis of a systematic literature review on application of artificial intelligence to analyze/process social media big data for efficient disaster management. During a systematic review process 68 publications were identified. Thereafter, we analyzed all the identified papers. From our analysis, we conclude that the most of the reviewed papers are on text and image classification and mostly convolutional neural networks have been employed for the classification.

Being capable of online learning in unknown stochastic environments, Tsetlin Automata (TA) have gained considerable interest. As a model of biological systems, teams of TA have been used for solving complex problems in a decentralized manner, with low computational complexity. For many domains, decentralized problem solving is an advantage, however, also may lead to coordination difficulties and unstable learning. To combat this negative effect, this paper proposes a novel TA coordination scheme designed for learning problems with continuous input and output. By saving and updating the best solution that has been chosen so far, we can avoid having the overall system being led astray by spurious erroneous actions. We organize this process hierarchically by a principal-teacherclass structure. We further propose a binary representation of continuous actions (coefficients). Each coefficient in the cost function is represented by 8 TA. TA teams at different classes produce different solutions. They are trained to find the global optimum with the help of their own best and the overall best solutions. The proposed algorithm is tested first with an artificial dataset and later used to forecast dengue haemorrhagic fever in the Philippines. Results of the novel procedure are compared with results from two traditional TA approaches. The training error of the novel TA scheme is lower approx. 50 and 62 times compared to the considered two traditional Tsetlin Automata approaches and testing error is approx. 31 and 21 times lower for the new scheme. These improvements not only highlight the effectiveness of the proposed scheme, but also the importance of old, simple, yet powerful concepts in the Artificial Intelligence techniques.

Diabetes is one of the worlds major health problems according to the World Health Organization. Recent surveys indicate that there is an increase in the number of diabetic patients resulting in an increase in serious complications such as heart attacks and deaths. Early diagnosis of diabetes, particularly of type 2 diabetes, is critical since it is vital for patients to get insulin treatments. However, diagnoses could be difficult especially in areas with few medical doctors. It is, therefore, a need for practical methods for the public for early detection and prevention with minimal intervention from medical professionals. A promising method for automated diagnosis is the use of artificial intelligence and in particular artificial neural networks. This paper presents an application of Multi-Layer Feed Forward Neural Networks (MLFNN) in diagnosing diabetes on publicly available Pima Indian Diabetes (PID) data set. A series of experiments are conducted on this data set with variation in learning algorithms, activation units, techniques to handle missing data and their impact on diagnosis accuracy is discussed. Finally, the results are compared with other states of art methods reported in the literature review. The achieved accuracy is 82.5% best of all related studies.

In this paper, we propose a hybrid model consisting of a Deep Convolutional feature extractor followed by a fast and accurate classifier, the Extreme Learning Machine, for the purpose of fire detection in images. The reason behind using such a model is that Deep CNNs used for image classification take a very long time to train. Even with pre-trained models, the fully connected layers need to be trained with backpropagation, which can be very slow. In contrast, we propose to employ the Extreme Learning Machine (ELM) as the final classifier trained on pre-trained Deep CNN feature extractor. We apply this hybrid model on the problem of fire detection in images. We use state of the art Deep CNNs: VGG16 and Resnet50 and replace the softmax classifier with the ELM classifier. For both the VGG16 and Resnet50, the number of fully connected layers is also reduced. Especially in VGG16, which has 3 fully connected layers of 4096 neurons each followed by a softmax classifier, we replace two of these with an ELM classifier. The difference in convergence rate between fine-tuning the fully connected layers of pre-trained models and training an ELM classifier are enormous, around 20x to 51x speed-up. Also, we show that using an ELM classifier increases the accuracy of the system by 2.8% to 7.1% depending on the CNN feature extractor. We also compare our hybrid architecture with another hybrid architecture, i.e. the CNN-SVM model. Using SVM as the classifier does improve accuracy compared to state-of-the-art deep CNNs. But our Deep CNN-ELM model is able to outperform the Deep CNN-SVM models. (Preliminary version of some of the results of this paper appear in “Deep Convolutional Neural Networks for Fire Detection in Images”, Springer Proceedings Engineering Applications of Neural Networks 2017 (EANN’17), Athens, Greece, 25–27 August).

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. This success is primarily due to the vast capabilities of convolutional neural networks, that can extract useful features from noisy and complex data. Games are excellent tools to test and push the boundaries of novel RL algorithms because they give valuable insight into how well an algorithm can perform in isolated environments without the real-life consequences. Real-time strategy games (RTS) is a genre that has tremendous complexity and challenges the player in short and long-term planning. There is much research that focuses on applied RL in RTS games, and novel advances are therefore anticipated in the not too distant future. However, there are to date few environments for testing RTS AIs. Environments in the literature are often either overly simplistic, such as microRTS, or complex and without the possibility for accelerated learning on consumer hardware like StarCraft II. This paper introduces the Deep RTS game environment for testing cutting-edge artificial intelligence algorithms for RTS games. Deep RTS is a high-performance RTS game made specifically for artificial intelligence research. It supports accelerated learning, meaning that it can learn at a magnitude of 50 000 times faster compared to existing RTS games. Deep RTS has a flexible configuration, enabling research in several different RTS scenarios, including partially observable state-spaces and map complexity. We show that Deep RTS lives up to our promises by comparing its performance with microRTS, ELF, and StarCraft II on high-end consumer hardware. Using Deep RTS, we show that a Deep Q-Network agent beats random-play agents over 70% of the time. Deep RTS is publicly available at https://github.com/cair/DeepRTS., Comment: Proceedings of the IEEE International Conference on Computational Intelligence and Games (CIG 2018)

Reinforcement learning has shown great potential in generalizing over raw sensory data using only a single neural network for value optimization. There are several challenges in the current state-of-the-art reinforcement learning algorithms that prevent them from converging towards the global optima. It is likely that the solution to these problems lies in short- and long-term planning, exploration and memory management for reinforcement learning algorithms. Games are often used to benchmark reinforcement learning algorithms as they provide a flexible, reproducible, and easy to control environment. Regardless, few games feature a state-space where results in exploration, memory, and planning are easily perceived. This paper presents The Dreaming Variational Autoencoder (DVAE), a neural network based generative modeling architecture for exploration in environments with sparse feedback. We further present Deep Maze, a novel and flexible maze engine that challenges DVAE in partial and fully-observable state-spaces, long-horizon tasks, and deterministic and stochastic problems. We show initial findings and encourage further work in reinforcement learning driven by generative exploration.

Chronic Obstructive Pulmonary Disease (COPD) patients need help in daily life situations as they are burdened with frequent risks of acute exacerbation and loss of control. An automated monitoring system could lead to timely treatments and avoid unnecessary hospital (re-)admissions and home visits by doctors or nurses. Therefore we present a Deep Artificial Neural Networks for approach prediction of exacerbations, particularly Feed-Forward Neural Networks (FFNN) for classification of COPD patients category and Long Short-Term Memory (LSTM), for early prediction of COPD exacerbations and subsequent triage. The FFNN and LSTM models are trained on data collected from remote monitoring of 94 patients through a real monitoring session and therefore represents realistic home monitoring situations. Most deep learning models require large datasets in order to predict with a high degree of accuracy. Our experiments show that with only 94 patients, the FFNN model is able to reproduce health condition provided by a medical doctor with an accuracy of 92.86% and the LSTM model able to predict COPD patients’ health conditions one-day ahead with an accuracy of 84.12%. Based on our results, we believe that our work will help the medical doctors and nurses in identifying patients with acute exacerbation in advance which can lead to better patient care and decision making, and hence reduction of costs.

Understanding of textual content, such as topic and intent recognition, is a critical part of chatbots, allowing the chatbot to provide relevant responses. Although successful in several narrow domains, the potential diversity of content in broader and more open domains renders traditional pattern recognition techniques inaccurate. In this paper, we propose a novel deep learning architecture for content recognition that consists of multiple levels of gated recurrent units (GRUs). The architecture is designed to capture complex sentence structure at multiple levels of abstraction, seeking content recognition for very wide domains, through a distributed scalable representation of content. To evaluate our architecture, we have compiled 10 years of questions and answers from a youth information service, \(200\ 083\) questions spanning a wide range of content, altogether 289 topics, involving law, health, and social issues. Despite the relatively open domain data set, our architecture is able to accurately categorize the 289 intents and topics. Indeed, it provides roughly an order of magnitude higher accuracy compared to more classical content recognition techniques, such as SVM, Naive Bayes, random forest, and K-nearest neighbor, which all seem to fail on this challenging open domain dataset.

Physics-based character animation offers an attractive alternative for traditional animations. However, it is often strenuous for a physics-based approach to incorporate active user control of different characters. In this paper, a neuroevolutionary approach is proposed using HyperNEAT to combine individually trained neural controllers to form a control strategy for a simulated eight-legged character, which is a previously untested character morphology for this algorithm. It is aimed to evaluate the robustness and responsiveness of the control strategy that changes the controllers based on simulated user inputs. The experiment result shows that HyperNEAT is able to evolve long walking controllers for this character. In addition, it also suggests a requirement for further refinement when operated in tandem.

Ant Colony Optimisation for classification has mostly been limited to rule based approaches where artificial ants walk on datasets in order to extract rules from the trends in the data, and hybrid approaches which attempt to boost the performance of existing classifiers through guided feature reductions or parameter optimisations. A recent notable example that is distinct from the mainstream approaches is PolyACO, which is a proof of concept polygon-based classifier that resorts to ant colony optimisation as a technique to create multi-edged polygons as class separators. Despite possessing some promise, PolyACO has some significant limitations, most notably, the fact of supporting classification of only two classes, including two features per class.This paper introduces PolyACO+, which is an extension of PolyACO in three significant ways: (1) PolyACO+ supports classifying multiple classes, (2) PolyACO+ supports polygons in multiple dimensions enabling classification with more than two features, and (3) PolyACO+ substantially reduces the training time compared to PolyACO by using the concept of multi-leveling. This paper empirically demonstrates that these updates improve the algorithm to such a degree that it becomes comparable to state-of-the-art techniques such as SVM, Neural Networks, and AntMiner+. nivå1