Algorithmic and human prediction of success in human collaboration from visual features

AbstractAs groups are increasingly taking over individual experts in many tasks, it is ever more important to understand the determinants of group success. In this paper, we study the patterns of group success in Escape The Room, a physical adventure game in which a group is tasked with escaping a maze by collectively solving a series of puzzles. We investigate (1) the characteristics of successful groups, and (2) how accurately humans and machines can spot them from a group photo. The relationship between these two questions is based on the hypothesis that the characteristics of successful groups are encoded by features that can be spotted in their photo. We analyze >43K group photos (one photo per group) taken after groups have completed the game—from which all explicit performance-signaling information has been removed. First, we find that groups that are larger, older and more gender but less age diverse are significantly more likely to escape. Second, we compare humans and off-the-shelf machine learning algorithms at predicting whether a group escaped or not based on the completion photo. We find that individual guesses by humans achieve 58.3% accuracy, better than random, but worse than machines which display 71.6% accuracy. When humans are trained to guess by observing only four labeled photos, their accuracy increases to 64%. However, training humans on more labeled examples (eight or twelve) leads to a slight, but statistically insignificant improvement in accuracy (67.4%). Humans in the best training condition perform on par with two, but worse than three out of the five machine learning algorithms we evaluated. Our work illustrates the potentials and the limitations of machine learning systems in evaluating group performance and identifying success factors based on sparse visual cues.

Download Full-text

Application of machine learning algorithms to predict permeability in tight sandstone formations

Nafta-Gaz ◽

10.18668/ng.2021.05.01 ◽

2021 ◽

Vol 77 (5) ◽

pp. 283-292

Author(s):

Tomasz Topór ◽

Keyword(s):

Machine Learning ◽

Oil And Gas ◽

Learning Algorithms ◽

Confining Pressure ◽

Machine Learning Algorithms ◽

Core Material ◽

Confining Stress ◽

Tight Sandstone ◽

Oil And Gas Exploration ◽

Better Than

The application of machine learning algorithms in petroleum geology has opened a new chapter in oil and gas exploration. Machine learning algorithms have been successfully used to predict crucial petrophysical properties when characterizing reservoirs. This study utilizes the concept of machine learning to predict permeability under confining stress conditions for samples from tight sandstone formations. The models were constructed using two machine learning algorithms of varying complexity (multiple linear regression [MLR] and random forests [RF]) and trained on a dataset that combined basic well information, basic petrophysical data, and rock type from a visual inspection of the core material. The RF algorithm underwent feature engineering to increase the number of predictors in the models. In order to check the training models’ robustness, 10-fold cross-validation was performed. The MLR and RF applications demonstrated that both algorithms can accurately predict permeability under constant confining pressure (R2 0.800 vs. 0.834). The RF accuracy was about 3% better than that of the MLR and about 6% better than the linear reference regression (LR) that utilized only porosity. Porosity was the most influential feature of the models’ performance. In the case of RF, the depth was also significant in the permeability predictions, which could be evidence of hidden interactions between the variables of porosity and depth. The local interpretation revealed the common features among outliers. Both the training and testing sets had moderate-low porosity (3–10%) and a lack of fractures. In the test set, calcite or quartz cementation also led to poor permeability predictions. The workflow that utilizes the tidymodels concept will be further applied in more complex examples to predict spatial petrophysical features from seismic attributes using various machine learning algorithms.

Download Full-text

A Review On The Relationship Between Human And Artificial Imagination With Their Implementations In Current Machine Learning Algorithms

10.1109/icccnt51525.2021.9579860 ◽

2021 ◽

Author(s):

Md. Abdullah-Al-Kafi ◽

Md. Fahad Hossain ◽

Sheak Rashed Haider Noori

Keyword(s):

Machine Learning ◽

Learning Algorithms ◽

Machine Learning Algorithms ◽

The Relationship

Download Full-text

Novel Privacy Preserving Non-Invasive Sensing-Based Diagnoses of Pneumonia Disease Leveraging Deep Network Model

Sensors ◽

10.3390/s22020461 ◽

2022 ◽

Vol 22 (2) ◽

pp. 461

Author(s):

Mujeeb Ur Rehman ◽

Arslan Shafique ◽

Kashif Hesham Khan ◽

Sohail Khalid ◽

Abdullah Alhumaidi Alotaibi ◽

...

Keyword(s):

Machine Learning ◽

Deep Learning ◽

Medical Records ◽

Learning Algorithms ◽

Machine Learning Algorithms ◽

X Ray ◽

Non Invasive ◽

Proposed Model ◽

Pneumonia Diagnosis ◽

Better Than

This article presents non-invasive sensing-based diagnoses of pneumonia disease, exploiting a deep learning model to make the technique non-invasive coupled with security preservation. Sensing and securing healthcare and medical images such as X-rays that can be used to diagnose viral diseases such as pneumonia is a challenging task for researchers. In the past few years, patients’ medical records have been shared using various wireless technologies. The wireless transmitted data are prone to attacks, resulting in the misuse of patients’ medical records. Therefore, it is important to secure medical data, which are in the form of images. The proposed work is divided into two sections: in the first section, primary data in the form of images are encrypted using the proposed technique based on chaos and convolution neural network. Furthermore, multiple chaotic maps are incorporated to create a random number generator, and the generated random sequence is used for pixel permutation and substitution. In the second part of the proposed work, a new technique for pneumonia diagnosis using deep learning, in which X-ray images are used as a dataset, is proposed. Several physiological features such as cough, fever, chest pain, flu, low energy, sweating, shaking, chills, shortness of breath, fatigue, loss of appetite, and headache and statistical features such as entropy, correlation, contrast dissimilarity, etc., are extracted from the X-ray images for the pneumonia diagnosis. Moreover, machine learning algorithms such as support vector machines, decision trees, random forests, and naive Bayes are also implemented for the proposed model and compared with the proposed CNN-based model. Furthermore, to improve the CNN-based proposed model, transfer learning and fine tuning are also incorporated. It is found that CNN performs better than other machine learning algorithms as the accuracy of the proposed work when using naive Bayes and CNN is 89% and 97%, respectively, which is also greater than the average accuracy of the existing schemes, which is 90%. Further, K-fold analysis and voting techniques are also incorporated to improve the accuracy of the proposed model. Different metrics such as entropy, correlation, contrast, and energy are used to gauge the performance of the proposed encryption technology, while precision, recall, F1 score, and support are used to evaluate the effectiveness of the proposed machine learning-based model for pneumonia diagnosis. The entropy and correlation of the proposed work are 7.999 and 0.0001, respectively, which reflects that the proposed encryption algorithm offers a higher security of the digital data. Moreover, a detailed comparison with the existing work is also made and reveals that both the proposed models work better than the existing work.

Download Full-text

Comparing Machine Learning Algorithms for Predicting ICU Admission and Mortality in COVID-19

10.1101/2020.11.20.20235598 ◽

2020 ◽

Author(s):

Sonu Subudhi ◽

Ashish Verma ◽

Ankit B. Patel ◽

C. Corey Hardin ◽

Melin J. Khandekar ◽

...

Keyword(s):

Machine Learning ◽

Clinical Decision Making ◽

Learning Algorithms ◽

Disease Outbreaks ◽

Clinical Decision ◽

Machine Learning Algorithms ◽

Healthcare Database ◽

Icu Admission ◽

Infectious Disease Outbreaks ◽

Better Than

AbstractAs predicting the trajectory of COVID-19 disease is challenging, machine learning models could assist physicians determine high-risk individuals. This study compares the performance of 18 machine learning algorithms for predicting ICU admission and mortality among COVID-19 patients. Using COVID-19 patient data from the Mass General Brigham (MGB) healthcare database, we developed and internally validated models using patients presenting to Emergency Department (ED) between March-April 2020 (n = 1144) and externally validated them using those individuals who encountered ED between May-August 2020 (n = 334). We show that ensemble-based models perform better than other model types at predicting both 5-day ICU admission and 28-day mortality from COVID-19. CRP, LDH, and procalcitonin levels were important for ICU admission models whereas eGFR <60 ml/min/1.73m2, ventilator use, and potassium levels were the most important variables for predicting mortality. Implementing such models would help in clinical decision-making for future COVID-19 and other infectious disease outbreaks.

Download Full-text

Optimising an FFQ Using a Machine Learning Pipeline to teach an Efficient Nutrient Intake Predictive Model

Nutrients ◽

10.3390/nu12123789 ◽

2020 ◽

Vol 12 (12) ◽

pp. 3789

Author(s):

Nina Reščič ◽

Tome Eftimov ◽

Barbara Koroušić Seljak ◽

Mitja Luštrek

Keyword(s):

Machine Learning ◽

Diet Quality ◽

Short Form ◽

Learning Algorithms ◽

Machine Learning Algorithms ◽

Paper Machine ◽

Food Frequency ◽

Promethee Method ◽

Dietary Monitoring ◽

Better Than

Food frequency questionnaires (FFQs) are the most commonly selected tools in nutrition monitoring, as they are inexpensive, easily implemented and provide useful information regarding dietary intake. They are usually carefully drafted by experts from nutritional and/or medical fields and can be validated by using other dietary monitoring techniques. FFQs can get very extensive, which could indicate that some of the questions are less significant than others and could be omitted without losing too much information. In this paper, machine learning is used to explore how reducing the number of questions affects the predicted nutrient values and diet quality score. The paper addresses the problem of removing redundant questions and finding the best subset of questions in the Extended Short Form Food Frequency Questionnaire (ESFFFQ), developed as part of the H2020 project WellCo. Eight common machine-learning algorithms were compared on different subsets of questions by using the PROMETHEE method, which compares methods and subsets via multiple performance measures. According to the results, for some of the targets, specifically sugar intake, fiber intake and protein intake, a smaller subset of questions are sufficient to predict diet quality scores. Additionally, for smaller subsets of questions, machine-learning algorithms generally perform better than statistical methods for predicting intake and diet quality scores. The proposed method could therefore be useful for finding the most informative subsets of questions in other FFQs as well. This could help experts develop FFQs that provide the necessary information and are not overbearing for those answering.

Download Full-text

Optimized Classification Predictions with a New Index Combining Machine Learning Algorithms

International Journal of Artificial Intelligence Tools ◽

10.1142/s0218213018500124 ◽

2018 ◽

Vol 27 (03) ◽

pp. 1850012 ◽

Cited By ~ 6

Author(s):

Androniki Tamvakis ◽

Christos-Nikolaos Anagnostopoulos ◽

George Tsirtsis ◽

Antonios D. Niros ◽

Sofie Spatharis

Keyword(s):

Machine Learning ◽

Learning Algorithms ◽

Individual Performance ◽

Optimal Combination ◽

Machine Learning Algorithms ◽

Combined Effect ◽

Different Types ◽

Open Issue ◽

The Relationship ◽

Selection Of

Voting is a commonly used ensemble method aiming to optimize classification predictions by combining results from individual base classifiers. However, the selection of appropriate classifiers to participate in voting algorithm is currently an open issue. In this study we developed a novel Dissimilarity-Performance (DP) index which incorporates two important criteria for the selection of base classifiers to participate in voting: their differential response in classification (dissimilarity) when combined in triads and their individual performance. To develop this empirical index we firstly used a range of different datasets to evaluate the relationship between voting results and measures of dissimilarity among classifiers of different types (rules, trees, lazy classifiers, functions and Bayes). Secondly, we computed the combined effect on voting performance of classifiers with different individual performance and/or diverse results in the voting performance. Our DP index was able to rank the classifier combinations according to their voting performance and thus to suggest the optimal combination. The proposed index is recommended for individual machine learning users as a preliminary tool to identify which classifiers to combine in order to achieve more accurate classification predictions avoiding computer intensive and time-consuming search.

Download Full-text

Using Machine Learning Algorithms to Evaluate the Relationship Between Air Quality and Temperature Change

10.1002/essoar.10501313.1 ◽

2019 ◽

Author(s):

Yuxi Jin

Keyword(s):

Machine Learning ◽

Air Quality ◽

Temperature Change ◽

Learning Algorithms ◽

Machine Learning Algorithms ◽

The Relationship

Download Full-text

Is Predicting Software Security Bugs Using Deep Learning Better Than the Traditional Machine Learning Algorithms?

2018 IEEE International Conference on Software Quality, Reliability and Security (QRS) ◽

10.1109/qrs.2018.00023 ◽

2018 ◽

Cited By ~ 1

Author(s):

Caesar Jude Clemente ◽

Fehmi Jaafar ◽

Yasir Malik

Keyword(s):

Machine Learning ◽

Deep Learning ◽

Software Security ◽

Learning Algorithms ◽

Machine Learning Algorithms ◽

Better Than

Download Full-text

EEG correlation at a distance: A re-analysis of two studies using a machine learning approach

F1000Research ◽

10.12688/f1000research.17613.2 ◽

2019 ◽

Vol 8 ◽

pp. 43 ◽

Cited By ~ 2

Author(s):

Marco Bilucaglia ◽

Luciano Pederzoli ◽

William Giroldini ◽

Elena Prati ◽

Patrizio Tressoldi

Keyword(s):

Machine Learning ◽

Learning Algorithms ◽

Machine Learning Algorithms ◽

Eeg Activity ◽

Linear Discriminant ◽

Machine Learning Approach ◽

Sensorial Stimulation ◽

The Relationship ◽

Electroencephalogram Eeg ◽

Linear Discriminant Classifier

Background: In this paper, data from two studies relative to the relationship between the electroencephalogram (EEG) activities of two isolated and physically separated subjects were re-analyzed using machine-learning algorithms. The first dataset comprises the data of 25 pairs of participants where one member of each pair was stimulated with a visual and an auditory 500 Hz signals of 1 second duration. The second dataset consisted of the data of 20 pairs of participants where one member of each pair received visual and auditory stimulation lasting 1 second duration with on-off modulation at 10, 12, and 14 Hz. Methods and Results: Applying a ‘linear discriminant classifier’ to the first dataset, it was possible to correctly classify 50.74% of the EEG activity of non-stimulated participants, correlated to the remote sensorial stimulation of the distant partner. In the second dataset, the percentage of correctly classified EEG activity in the non-stimulated partners was 51.17%, 50.45% and 51.91%, respectively, for the 10, 12, and 14 Hz stimulations, with respect the condition of no stimulation in the distant partner. Conclusions: The analysis of EEG activity using machine-learning algorithms has produced advances in the study of the connection between the EEG activities of the stimulated partner and the isolated distant partner, opening new insight into the possibility to devise practical application for non-conventional “mental telecommunications” between physically and sensorially separated participants.

Download Full-text