Articles tagged with Machine Learning
A new UT Dallas study derived optimal policies and data-driven techniques for firms to learn about demand and adjust capacity. The study's main finding is that production managers need to maintain a careful balance between observing the demand and changing the capacity.
Swiss researchers use neural networks to challenge the resolution limit of telescopes, recovering features that were previously invisible. The technique, inspired by a generative adversarial network, achieves better results than previous methods, such as deconvolution, and has vast potential for future astronomical observations.
A Northwestern University and Los Alamos National Laboratory team developed a novel workflow to design new materials with useful electronic properties. By combining machine learning and density functional theory calculations, they created design guidelines for ferroelectricity and piezoelectricity.
A new project by Upside Energy and Heriot-Watt University aims to use machine learning and distributed AI to manage storage assets and provide real-time energy reserves. The goal is to relieve stress on the grid and reduce reliance on traditional power stations.
Researchers have identified nearly two-dozen solid electrolytes that could replace volatile liquids in smartphones and laptops. The AI-powered approach allows for rapid screening of materials, identifying the most promising candidates for further study.
Researchers are preparing to tackle an onslaught of petabytes of complex data from sophisticated experiments, including CERN's Large Hadron Collider. To keep up with the challenge, experts propose developing exascale supercomputers and smarter networks, as well as reengineering software to adapt to future hardware developments.
A team of researchers from the University of Toronto has developed a novel machine learning approach to determine whether planetary systems are stable or not. This method is 1,000 times faster than traditional methods and can provide valuable information about exoplanets, including their mass and orbital eccentricity.
Researchers at U of T Engineering developed an AI algorithm that learns directly from human instructions, exceeding conventional training methods by 160% and outperforming its own training by 9%. The algorithm's potential lies in applying heuristic training to fields like medicine and transportation.
Researchers developed a machine learning classifier to discover membrane-active peptides with diverse sequences. The approach identified new peptides with broad biomedical implications, including immunotherapy and anticancer therapeutics.
Researchers at Numenta compared their biologically-derived HTM sequence memory to traditional machine learning algorithms, demonstrating comparable prediction accuracy. The new paper highlights the algorithm's properties, including continuous online learning and robustness to sensor noise, making it ideal for streaming data applications.
Researchers developed a new AI system that uses web search to extract data from plain text, outperforming traditional machine learning methods by up to 10%. The system learns to generate search queries and gauge relevance, then extracts relevant data from online texts.
Researchers at UTA are developing an AI system that assesses children's executive function skills, recognizing patterns of inattention and hyperactivity. The system provides recommendations for effective intervention and monitoring progress over time.
A new data-cleaning tool called ActiveClean analyzes a user's prediction model to identify mistakes and update the model as it works. By minimizing human error, ActiveClean improves model accuracy and reduces statistical biases, making it an essential tool for building better prediction models.
Researchers have discovered a way for machines to learn about natural or artificial systems by observing them, eliminating the need for prior knowledge. This breakthrough could lead to advances in technology, including predictive human behavior and algorithm development for detecting abnormalities.
Researchers developed a machine learning system that detects differences in accelerometer data between individuals with muscle tension dysphonia and controls. The system improved after receiving voice therapy, suggesting potential for wearable devices providing real-time feedback.
Researchers developed a method to identify impoverished areas using high-resolution satellite imagery and machine learning. The approach outperformed existing methods and could be used to map poverty worldwide at low cost.
Researchers used direct neural recordings and brain stimulation to study visual word form area's role in reading. They found that this area codes knowledge about learned visual words, enabling accurate discrimination of similar words.
Researchers led by Josef van Genabith are working on two EU-funded projects to improve automatic machine translation, particularly for complex languages like Latvian and Czech. They aim to use 'deep learning' to recognize patterns in large text datasets and learn from them.
Researchers have developed an AI-powered MRI technique that can detect early forms of dementia, including mild cognitive impairment and Alzheimer's disease. The technique uses machine learning to analyze perfusion maps created by arterial spin labeling (ASL) imaging, with high accuracy in distinguishing between patients.
Computer vision systems can now learn to recognize objects they have never seen before by analyzing word use and contextualization, reducing the need for thousands of labeled images. This new learning paradigm, called semi-supervised vocabulary-informed learning, was developed by Disney Researchers using a large dataset of English words.
Automated camera system improves sports broadcasts by learning from human operators and achieving smoother footage without jerkiness. The system uses a new approach called imitation learning, which repeats multiple times and analyzes deviations to learn from human mistakes.
A team of researchers led by Sridhar Mahadevan is applying deep learning methods to analyze large amounts of scientific data from Mars. They hope to develop a practical tool for handling vast amounts of data created by various scientific instruments, including the NASA Curiosity rover.
Scientists from Insilico Medicine used deep neural networks to predict therapeutic use of large numbers of drugs from gene expression data, achieving 54.6% accuracy in class prediction. The study also found that many misclassified drugs had dual use, suggesting potential for drug repurposing.
Insilico Medicine presents research on applying deep learning to biomarker development and cosmetics applications at INNOCOS World Beauty Innovation Summit. The company's app RYNKL evaluates anti-aging interventions using machine learning methods, minimizing animal testing.
A University of Washington team developed a highly capable five-fingered robot hand that can perform dexterous manipulation and learn from its own experience. The hand uses machine learning algorithms to model physics and plan actions, allowing it to adapt to new tasks without human intervention.
Researchers explore machine learning's potential to enhance plastic surgery with algorithms for predicting burn healing times and suggesting reconstructive approaches. The field also holds promise for improving microsurgery, craniofacial surgery, hand and peripheral nerve surgery, and aesthetic surgery outcomes.
Researchers at Carnegie Mellon University have identified specific neural systems used to encode new scientific concepts. The study shows that the brain repurposes existing neural structures to form new knowledge, enabling humans to learn abstract ideas.
Researchers at Numenta Inc. have published a new theory on how networks of neurons in the neocortex learn sequences, providing a breakthrough in understanding neural circuits.
Researchers at Tel Aviv University have developed a cutting-edge solution for radiologists using Deep Learning technologies. The lab has created tools to facilitate computer-assisted diagnosis of X-rays, CTs and MRIs, freeing up time for complex cases.
A novel approach, called machine unlearning, enables the removal of data without retraining a computer learning system from scratch. This method is crucial for increasing security and protecting user privacy, especially in the face of cyber-attacks and data breaches.
Researchers developed a web-based platform using artificial neural networks to answer crossword clues more accurately than existing products. The system can understand words, phrases, and sentences by leveraging definitions in dictionaries and Wikipedia.
Carnegie Mellon University's Jing Lei and Ryan Tibshirani have been awarded NSF CAREER grants for their cutting-edge research projects in large-scale data analysis and nonparametric estimation. Their work aims to advance statistical inference with complex high-dimensional data.
Todd Gureckis, NYU associate professor, awarded PECASE for pioneering research on human cognition and machine learning. The award recognizes his innovative work in comparing human intelligence to intelligent algorithms.
Researchers from MIPT and MSU created a computer model predicting agrochemical activity using machine learning and Kohonen self-organizing maps. The model showed high predictive power, accurately classifying molecules into pesticides or plant growth regulators with 87% accuracy.
A new robotically driven experimentation system has been developed to determine the effects of a large number of drugs on many proteins, reducing the number of necessary experiments by 70%. The model uses an active learning approach to identify patterns and make predictions about unmeasured experiments with high accuracy.
Researchers at University of California, Riverside developed an energy management system that improves plug-in hybrid efficiency by 12 percent. The system uses machine learning and real-time data to optimize energy consumption and reduce greenhouse gas emissions.
Researchers will use insights from neural circuitry and learning methods to create more human-like computer vision and machine learning algorithms. The CMU-led team aims to unlock the brain's secret algorithms in learning and inference with a massive database of neural activity.
Researchers at Harvard are using a $28 million grant to study the brain's visual cortex in unprecedented detail and map its connections. The goal is to inspire better computer algorithms for learning and pattern recognition, enabling computers to outperform humans in recognizing patterns from limited data inputs.
A team of researchers is exploring the use of the NAO robot in a new approach to pediatric rehabilitation based on social interaction between robots and humans. The robot can read moods, recognize family members, and learn preferences, providing personalized interventions for children with motor disabilities.
InSilico Medicine presents recent advances in applying signaling pathway activation analysis and deep learning to drug discovery and age-related diseases. The company's mission is to extend healthy human longevity through faster and more effective diagnostics and cures.
Researchers at Carnegie Mellon University have developed a high-throughput, machine-learning tool to analyze synaptic density in the brain. This allows them to identify synapses from an entire cortical region and gain insights into how synaptic properties change during development and learning.
Researchers found that humans can recognize objects even when only a small portion is visible, and validated an algorithm to explain human learning. The algorithm can be used for machine learning, data analysis, and computer vision to improve performance.
Researchers developed a Bayesian Program Learning framework that captures human learning abilities, allowing computers to recognize and generate new visual concepts. The algorithm achieved impressive results in visual Turing tests, with only 25% of judges performing better than chance.
Researchers developed a computer model called Bayesian program learning framework (BPL) that captures humans' unique ability to learn new concepts from a single example. The BPL model achieved human-level performance on challenging concept learning tasks, outperforming recent deep learning approaches.
Researchers at the University of Washington have created a new probabilistic model that allows robots to learn new skills by watching people and imitating them. The team combined child development research with machine learning approaches, inspired by infants' ability to infer adult intentions through self-exploration.
The study aims to develop machine learning models that mimic the toddlers' ability to recognize objects, leveraging 500 hours of video and 54 million images from over 100 children. The research could lead to more sophisticated digital object-recognition technology.
Researchers at UMass Amherst aim to develop a reliable, predictive computational framework for designing better-performing materials with reduced development costs. The new approach will address challenges in handling complex systems with millions of variables, rare events, and multi-scale features.
Researchers at the University of Wisconsin-Madison are developing a new approach called machine teaching, which uses sophisticated mathematics to model human learners and devise the best possible lessons. This method has immense potential to impact education by providing optimal, personalized lessons for students in various fields.
Researchers developed a tool enabling art directors to control computer programs using verbal descriptors like 'silky' and 'wrinkly'. The system improved the process of creating garments with desired properties, reducing laborious tweaking of technical parameters.
Researchers at University College London developed a novel AI approach to predict fine wine prices more accurately, outperforming traditional methods by 15% on average. The new method uses machine learning to learn relevant information from data and improve predictive accuracy.
DeepBind uses deep learning to analyze protein-DNA/RNA binding and detect mutations that can disrupt cellular processes. The tool provides new information on disruptions in mutations tied to cancers, haemophilia, and familial hypercholesterolemia.
Recent progress in AI has driven significant advancements in machine learning, language processing, and human-like computer programs. Researchers discuss potential implications for humankind, including challenges to privacy, equality, and the public good.
Researchers developed a dropout-prediction model that uses data from one course offering to predict stopout in the next. The model achieved fairly accurate predictions and showed promise, particularly when incorporating additional variables like weekend study habits. Ongoing work aims to refine the model for improved accuracy.
A new study uses machine learning to pinpoint rodent species effective disease reservoirs and identify geographic hotspots vulnerable to emerging diseases. The research provides a basis for targeted surveillance efforts, highlighting the importance of collaboration with experts on the ground.
Researchers at MIT have developed a probabilistic programming language called Picture that can solve computer-vision tasks using short programs. The new system, which is competitive with conventional systems, has been shown to improve error rates on certain tasks, such as human pose estimation.
Disney Research has developed a robotic camera system that can learn from human operators to better frame shots of a basketball game. The system uses machine learning algorithms to recognize the relationship between player locations and corresponding camera configurations.
A new automated method developed by Disney Research uses AI to select and order photos in a way that makes narrative sense, telling a compelling story. The system learns principles of selecting and ordering photos from large collections and can customize the process for individual preferences.
A new system developed at MIT enables pattern-recognition systems to distill what they learn into simple examples, which humans can use to make better decisions. In experiments, human subjects using the system outperformed those using a similar algorithm by over 20 percent.
Researchers at Carnegie Mellon University used fMRI scans to analyze brain activity while eight people read a chapter of Harry Potter. The result was the first integrated computational model of reading, identifying which parts of the brain are responsible for parsing sentences and determining word meaning.
A team of scientists developed a mathematical model that explains how the brain maintains stability during learning, resolving a decades-old paradox. The model suggests that fast and slow changes in neuronal networks work together to achieve sensitivity and stability.