Collaborating Authors


Interpretable and unsupervised phase classification


Fully automated classification methods that provide direct physical insights into phase diagrams are of current interest. Interpretable, i.e., fully explainable, methods are desired for which we understand why they yield a given phase classification. Ideally, phase classification methods should also be unsupervised. That is, they should not require prior labeling or knowledge of the phases of matter to be characterized. Here, we demonstrate an unsupervised machine-learning method for phase classification, which is rendered interpretable via an analytical derivation of the functional relationship between its optimal predictions and the input data. Based on these findings, we propose and apply an alternative, physically-motivated, data-driven scheme, which relies on the difference between mean input features. This mean-based method does not rely on any predictive model and is thus computationally cheap and directly explainable. As an example, we consider the physically rich ground-state phase diagram of the spinless Falicov-Kimball model.

Men with longer features and larger eyes are perceived as more promiscuous, study finds

Daily Mail - Science & tech

Men with long facial features and large eyes, and women with slim faces and small eyes are percieved as more promiscuous, a new study has revealed. However, this perception only rings true for men, and not for women, according to the researchers. In the study, experts in Australia asked heterosexual men and women about their levels of'sociosexuality' – the willingness to engage in sexual activity outside of a committed relationship, also known as casual sex. The participants also had their photos taken and shown to other participants of the opposite sex, so they could judge, based on looks alone, if they had an interest in sociosexuality. Men who were open to casual sex typically had longer faces, higher foreheads, longer noses and larger eyes, the team found.

Total Nitrogen Estimation in Agricultural Soils via Aerial Multispectral Imaging and LIBS


Measuring soil health indicators is an important and challenging task that affects farmers' decisions on timing, placement, and quantity of fertilizers applied in the farms. Most existing methods to measure soil health indicators (SHIs) are in-lab wet chemistry or spectroscopy-based methods, which require significant human input and effort, time-consuming, costly, and are low-throughput in nature. To address this challenge, we develop an artificial intelligence (AI)-driven near real-time unmanned aerial vehicle (UAV)-based multispectral sensing (UMS) solution to estimate total nitrogen (TN) of the soil, an important macro-nutrient or SHI that directly affects the crop health. Accurate prediction of soil TN can significantly increase crop yield through informed decision making on the timing of seed planting, and fertilizer quantity and timing. We train two machine learning models including multi-layer perceptron and support vector machine to predict the soil nitrogen using a suite of data classes including multispectral characteristics of the soil and crops in red, near-infrared, and green spectral bands, computed vegetation indices, and environmental variables including air temperature and relative humidity.

Education in the time of Covid-19, How can AI help?


Personalized Learning: The millions of people who use services like Amazon and Spotify have firsthand experience with personally curated choices. Millions if not billions have been invested in personalization algorithms with commercial objectives. Similar algorithms adapted for education could have an even greater benefit to society to deliver personalized learning experiences. Adaptive Learning: If personalized learning's objective is to present the individual with a learning path based on their learning history, behavior, and preferences then adaptive learning's objective is to optimize how much time and energy it takes to move through that path while mastering the learning goal. Adaptive learning detects whether the learner is struggling with a topic and presents more material on the topic in complementary ways helping the learner understand it better.

Market for Emotion Recognition Projected to Grow as Some Question Science - AI Trends


The emotion recognition software segment is projected to grow dramatically in coming years, spelling success for companies that have established a beachhead in the market, while causing some who are skeptical about its accuracy and fairness to raise red flags. The global emotion detection and recognition market is projected to grow to $37.1 billion by 2026, up from an estimated $19.5 billion in 2020, according to a recent report from MarketsandMarkets. North America is home to the largest market. Software suppliers covered in the report include: NEC Global (Japan), IBM (US), Intel (US), Microsoft (US), Apple (US), Gesturetek (Canada), Noldus Technology (Netherlands), Google (US), Tobii (Sweden), Cognitec Systems (Germany), Cipia Vision Ltd (Formerly Eyesight Technologies) (Israel), iMotions (Denmark), Numenta (US), Elliptic Labs (Norway), Kairos (US), PointGrab (US), Affectiva (US), nViso (Switzerland), Beyond Verbal (Israel), Sightcorp (Holland), Crowd Emotion (UK), Eyeris (US), Sentiance (Belgium), Sony Depthsense (Belgium), Ayonix (Japan), and Pyreos (UK). Some question whether emotion recognition software is effective, and whether its use is ethical.

51+ Data Sets for Beginner Data Science and Machine Learning Projects


Description -- This database, updated daily, contains ads that ran on Facebook and were submitted by thousands of ProPublica users from around the world. We asked our readers to install browser extensions that automatically collected advertisements on their Facebook pages and sent them to our servers. We then used a machine learning classifier to identify which ads were likely political and included them in this dataset.

StockBabble: A Conversational Financial Agent to support Stock Market Investors Artificial Intelligence

We introduce StockBabble, a conversational agent designed to support understanding and engagement with the stock market. StockBabble's value and novelty is in its ability to empower retail investors -- many of which may be new to investing -- and supplement their informational needs using a user-friendly agent. Users have the ability to query information on companies to retrieve a general and financial overview of a stock, including accessing the latest news and trading recommendations. They can also request charts which contain live prices and technical investment indicators, and add shares to a personal portfolio to allow performance monitoring over time. To evaluate our agent's potential, we conducted a user study with 15 participants. In total, 73% (11/15) of respondents said that they felt more confident in investing after using StockBabble, and all 15 would consider recommending it to others. These results are encouraging and suggest a wider appeal for such agents. Moreover, we believe this research can help to inform the design and development of future intelligent, financial personal assistants.

RCURRENCY: Live Digital Asset Trading Using a Recurrent Neural Network-based Forecasting System Artificial Intelligence

Consistent alpha generation, i.e., maintaining an edge over the market, underpins the ability of asset traders to reliably generate profits. Technical indicators and trading strategies are commonly used tools to determine when to buy/hold/sell assets, yet these are limited by the fact that they operate on known values. Over the past decades, multiple studies have investigated the potential of artificial intelligence in stock trading in conventional markets, with some success. In this paper, we present RCURRENCY, an RNN-based trading engine to predict data in the highly volatile digital asset market which is able to successfully manage an asset portfolio in a live environment. By combining asset value prediction and conventional trading tools, RCURRENCY determines whether to buy, hold or sell digital currencies at a given point in time. Experimental results show that, given the data of an interval $t$, a prediction with an error of less than 0.5\% of the data at the subsequent interval $t+1$ can be obtained. Evaluation of the system through backtesting shows that RCURRENCY can be used to successfully not only maintain a stable portfolio of digital assets in a simulated live environment using real historical trading data but even increase the portfolio value over time.

Towards Learning to Play Piano with Dexterous Hands and Touch Machine Learning

The virtuoso plays the piano with passion, poetry and extraordinary technical ability. As Liszt said (a virtuoso)must call up scent and blossom, and breathe the breath of life. The strongest robots that can play a piano are based on a combination of specialized robot hands/piano and hardcoded planning algorithms. In contrast to that, in this paper, we demonstrate how an agent can learn directly from machine-readable music score to play the piano with dexterous hands on a simulated piano using reinforcement learning (RL) from scratch. We demonstrate the RL agents can not only find the correct key position but also deal with various rhythmic, volume and fingering, requirements. We achieve this by using a touch-augmented reward and a novel curriculum of tasks. We conclude by carefully studying the important aspects to enable such learning algorithms and that can potentially shed light on future research in this direction.

Machine Learning Based Prediction of Future Stress Events in a Driving Scenario Artificial Intelligence

This paper presents a model for predicting a driver's stress level up to one minute in advance. Successfully predicting future stress would allow stress mitigation to begin before the subject becomes stressed, reducing or possibly avoiding the performance penalties of stress. The proposed model takes features extracted from Galvanic Skin Response (GSR) signals on the foot and hand and Respiration and Electrocardiogram (ECG) signals from the chest of the driver. The data used to train the model was retrieved from an existing database and then processed to create statistical and frequency features. A total of 42 features were extracted from the data and then expanded into a total of 252 features by grouping the data and taking six statistical measurements of each group for each feature. A Random Forest Classifier was trained and evaluated using a leave-one-subject-out testing approach. The model achieved 94% average accuracy on the test data. Results indicate that the model performs well and could be used as part of a vehicle stress prevention system.