A data-driven robust optimization algorithm for black-box cases: An application to hyper-parameter optimization of machine learning algorithms

2021 ◽  
pp. 107581
Author(s):  
Farshad Seifi ◽  
Mohammad Javad Azizi ◽  
Seyed Taghi Akhavan Niaki
Keyword(s):  
Machine Learning ◽  
Robust Optimization ◽  
Parameter Optimization ◽  
Optimization Algorithm ◽  
Learning Algorithms ◽  
Black Box ◽  
Machine Learning Algorithms ◽  
Data Driven

scholarly journals CoRg: Commonsense Reasoning Using a Theorem Prover and Machine Learning

10.29007/lt5p ◽  
2019 ◽  
Author(s):  
Sophie Siebert ◽  
Frieder Stolzenburg
Keyword(s):  
Machine Learning ◽  
Question Answering ◽  
Learning Algorithms ◽  
Knowledge Bases ◽  
Black Box ◽  
Machine Learning Algorithms ◽  
Theorem Prover ◽  
Commonsense Reasoning ◽  
Probable Answer ◽  
Everyday Task

Commonsense reasoning is an everyday task that is intuitive for humans but hard to implement for computers. It requires large knowledge bases to get the required data from, although this data is still incomplete or even inconsistent. While machine learning algorithms perform rather well on these tasks, the reasoning process remains a black box. To close this gap, our system CoRg aims to build an explainable and well-performing system, which consists of both an explainable deductive derivation process and a machine learning part. We conduct our experiments on the Copa question-answering benchmark using the ontologies WordNet, Adimen-SUMO, and ConceptNet. The knowledge is fed into the theorem prover Hyper and in the end the conducted models will be analyzed using machine learning algorithms, to derive the most probable answer.

scholarly journals An Intelligent and Data-Driven Mobile Platform for Youth Volunteer Management using Machine Learning and Predictive Analytics

2020 ◽  
Author(s):  
Alyssa Huang ◽  
Yu Sun
Keyword(s):  
Machine Learning ◽  
High School ◽  
High School Students ◽  
Predictive Analytics ◽  
Learning Algorithms ◽  
Real Life ◽  
Machine Learning Algorithms ◽  
Data Driven ◽  
School Students ◽  
Best Fit

Volunteering is very important to high school students because it not only allows the teens to apply the knowledge and skills they have acquired to real-life scenarios, but it also enables them to make an association between helping others and their own joy of fulfillment. Choosing the right volunteering opportunities to work on can influence how the teens interact with that cause and how well they can serve the community through their volunteering services. However, high school students who look for volunteer opportunities often do not have enough information about the opportunities around them, so they tend to take whatever opportunity that comes across. On the other hand, as organizations who look for volunteers usually lack effective ways to evaluate and select the volunteers that best fit the jobs, they will just take volunteers on a first-come, firstserve basis. Therefore, there is a need to build a platform that serves as a bridge to connect the volunteers and the organizations that offer volunteer opportunities. In this paper, we focus on creating an intelligent platform that can effectively evaluate volunteer performance and predict best-fit volunteer opportunities by using machine learning algorithms to study 1) the correlation between volunteer profiles (e.g. demographics, preferred jobs, talents, previous volunteering events, etc.) and predictive volunteer performance in specific events and 2) the correlation between volunteer profiles and future volunteer opportunities. Two highest-scoring machine learning algorithms are proposed to make predictions on volunteer performance and event recommendations. We demonstrate that the two highest-scoring algorithms are able to make the best prediction for each query. Alongside the practice with the algorithms, a mobile application, which can run on both iPhone and Android platforms is also created to provide a very convenient and effective way for the volunteers and event supervisors to plan and manage their volunteer activities. As a result of this research, volunteers and organizations that look for volunteers can both benefit from this data-driven platform for a more positive overall experience.

scholarly journals Optimizing Fractional Compositions To Achieve Extraordinary Properties

2021 ◽  
Author(s):  
Andrew Falkowski ◽  
Steven Kauwe ◽  
Taylor Sparks
Keyword(s):  
Machine Learning ◽  
Loss Function ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
Inverse Design ◽  
Data Driven ◽  
Model Parameters ◽  
Chemical Systems ◽  
Target Property ◽  
Successful Candidate

Traditional, data-driven materials discovery involves screening chemical systems with machine learning algorithms and selecting candidates that excel in a target property. The number of screening candidates grows infinitely large as the fractional resolution of compositions the number of included elements increases. The computational infeasibility and probability of overlooking a successful candidate grow likewise. Our approach shifts the optimization focus from model parameters to the fractions of each element in a composition. Using a pretrained network, CrabNet, and writing a custom loss function to govern a vector of element fractions, compositions can be optimized such that a predicted property is maximized or minimized. Single and multi-property optimization examples are presented that highlight the capabilities and robustness of this approach to inverse design.

scholarly journals Using country-level variables to classify countries according to the number of confirmed COVID-19 cases: An unsupervised machine learning approach

2020 ◽  
Vol 5 ◽  
pp. 56 ◽  
Author(s):  
Rodrigo M. Carrillo-Larco ◽  
Manuel Castillo-Cara
Keyword(s):  
Machine Learning ◽  
Case Fatality Rate ◽  
Learning Algorithms ◽  
Case Fatality ◽  
Machine Learning Algorithms ◽  
Data Driven ◽  
Mortality Data ◽  
Fatality Rate ◽  
Unsupervised Machine Learning ◽  
Country Level

Background: The COVID-19 pandemic has attracted the attention of researchers and clinicians whom have provided evidence about risk factors and clinical outcomes. Research on the COVID-19 pandemic benefiting from open-access data and machine learning algorithms is still scarce yet can produce relevant and pragmatic information. With country-level pre-COVID-19-pandemic variables, we aimed to cluster countries in groups with shared profiles of the COVID-19 pandemic. Methods: Unsupervised machine learning algorithms (k-means) were used to define data-driven clusters of countries; the algorithm was informed by disease prevalence estimates, metrics of air pollution, socio-economic status and health system coverage. Using the one-way ANOVA test, we compared the clusters in terms of number of confirmed COVID-19 cases, number of deaths, case fatality rate and order in which the country reported the first case. Results: The model to define the clusters was developed with 155 countries. The model with three principal component analysis parameters and five or six clusters showed the best ability to group countries in relevant sets. There was strong evidence that the model with five or six clusters could stratify countries according to the number of confirmed COVID-19 cases (p<0.001). However, the model could not stratify countries in terms of number of deaths or case fatality rate. Conclusions: A simple data-driven approach using available global information before the COVID-19 pandemic, seemed able to classify countries in terms of the number of confirmed COVID-19 cases. The model was not able to stratify countries based on COVID-19 mortality data.

What sound sources trigger misophonia? Not just chewing and breathing

2020 ◽  
Author(s):  
Heather Hansen ◽  
Andrew B. Leber ◽  
Zeynep M. Saygin
Keyword(s):  
Machine Learning ◽  
Learning Algorithms ◽  
Individual Variability ◽  
Machine Learning Algorithms ◽  
Data Driven ◽  
Sound Sources ◽  
Environmental Sounds

Misophonia is a condition that affects approximately one out of five individuals, yet is relatively understudied. Misophonia involves an aversion to particular environmental sounds; previous work focuses primarily on aversions to human mouth and nose sounds (e.g., chewing, sniffling), but there exists considerable individual variability in sounds reported as bothersome, warranting an objective and data-driven investigation. What types of sounds trigger aversive reactions in misophonia, and can we identify individuals as having misophonia based on their aversiveness ratings to certain sounds? Experiment 1 asked whether human-produced oral/nasal sounds were more aversive than human-produced non-oral/nasal sounds and nonhuman/nature sounds. Experiment 2 replicated these findings using a large public sound bank and different participants, and additionally asked whether machine-learning algorithms could predict the presence and severity of misophonia. Results showed that sounds from all three source categories – not just oral/nasal sounds – were rated as significantly more aversive to individuals with misophonia compared to controls; further, modeling all source categories classified misophonia with 89% accuracy and predicted severity of misophonia with a correlation of 0.75. This work suggests that misophonia should be conceptualized as more than an aversion to human oral/nasal sounds, which has implications for future diagnostics and experimental consistency moving forward.

Sign in / Sign up

Export Citation Format

Share Document