Machine Learning News and Research

RSS

Machine learning is a subfield of artificial intelligence that focuses on developing algorithms and models capable of automatically learning and making predictions or decisions from data without being explicitly programmed. It involves training models on labeled datasets to recognize patterns and make accurate predictions or classifications in new, unseen data.

When AI Judges AI, Bias Becomes a Hidden Threat

As AI models become more advanced, their errors become more similar, undermining AI oversight mechanisms. A new metric, CAPA, reveals that AI judges and AI-assisted training reinforce systemic biases instead of mitigating them.

9 Feb 2025

AI Security Education Gets a Boost to Combat Growing Cyber Threats

Researchers from NJIT, Rutgers, and Temple University are developing AI security education programs to address adversarial machine learning threats, aiming to equip future engineers with robust defense strategies.

9 Feb 2025

AI Predicts Life-Saving Drug for Rare Disease, Patient Enters Remission

Researchers used an AI-driven approach to identify adalimumab as a potential life-saving treatment for idiopathic multicentric Castleman’s disease (iMCD), successfully reversing a terminal case. The breakthrough demonstrates AI’s potential for drug repurposing in rare diseases.

5 Feb 2025

AI and Quantum Geometry Drive Quest for Room-Temperature Superconductors

Researchers from Rice University and an international team have secured multimillion-dollar funding to develop next-generation superconductors using AI and quantum geometry, aiming for room-temperature superconductivity by 2033.

4 Feb 2025

MIT’s New AI Compiler Speeds Up Machine Learning by 30x Using Data Redundancies

MIT researchers developed an automated system that optimizes deep-learning models by leveraging both sparsity and symmetry in data, significantly reducing computational costs and energy use.

3 Feb 2025

AI’s Hidden Bias: Study Finds ChatGPT Skews Left in Text and Images

ChatGPT exhibits a left-leaning bias in text and image outputs, with refusals to generate certain right-leaning perspectives, raising concerns about AI fairness and societal impact.

3 Feb 2025

AI Risks Escalate as Advanced Models Gain Power and Autonomy

AI safety experts warn that rapid advancements in general-purpose AI pose increasing risks, from cyber threats to systemic economic impacts, urging better safeguards and international oversight.

2 Feb 2025

AI Safety Showdown: DeepSeek-R1 Falls Behind OpenAI’s o3-mini

DeepSeek-R1 is nearly ten times less safe than OpenAI’s o3-mini, with higher rates of unsafe responses, particularly in financial crime, violence, and hate speech categories. OpenAI’s policy safeguards play a crucial role in preventing harmful outputs.

2 Feb 2025

AI Enhances Electrospray Deposition for Advanced Manufacturing

Researchers at Binghamton University and the University at Buffalo are using AI and computational modeling to refine electrospray deposition, a technique for producing ultra-thin polymer films with applications in electronics and healthcare.

30 Jan 2025

AI-Driven Breakthrough Enhances LIGO’s Gravitational Wave Detection by Reducing Noise

Scientists at the University of California, Riverside, have developed an unsupervised machine learning tool that identifies patterns in LIGO's environmental data, reducing noise and improving gravitational wave detection.

30 Jan 2025

Explainable AI Boosts Wind Power Forecasting for Smarter Energy Grids

Researchers at EPFL's WiRE Lab have integrated explainable AI (XAI) into wind power forecasting models, improving transparency and reliability in predicting wind energy generation. Their study shows that XAI can identify key input variables, reducing uncertainty and making wind power more competitive in the energy market.

29 Jan 2025

AI Transforms Fuel Cell Materials Discovery, Boosting Efficiency and Innovation

AI is revolutionizing fuel cell research by accelerating material discovery, optimizing system parameters, and overcoming traditional trial-and-error limitations.

28 Jan 2025

Hackers vs. Defenders: How AI Is Changing the Cybersecurity Game

Cyber risk is a top concern for organizations, with AI playing a dual role in both cyber defense and cyber-attacks. The Gordon-Loeb Model provides a cost-benefit framework for optimizing cybersecurity investments.

28 Jan 2025

Smarter Solar: AI Finds High-Efficiency Materials Faster Than Ever

Researchers used AI to analyze a million potential molecules for perovskite solar cells, achieving a breakthrough with just 150 targeted experiments, improving efficiency to 26.2%.

23 Jan 2025

Predicting the Unpredictable: The New Era of Reservoir Computing

Researchers introduced a novel reservoir computing framework with a generalized readout, leveraging higher-order nonlinearities to enhance accuracy and robustness in predicting chaotic systems.

16 Jan 2025

AI Unlocks Catalytic Secrets with Infrared Spectroscopy

Research integrates infrared spectroscopy with machine learning to map spectroscopic fingerprints to atomistic structures, enabling precise monitoring of C–C coupling processes on catalytic surfaces. The model demonstrates broad applicability, including on doped Cu surfaces, advancing AI-driven chemical reaction insights.

16 Jan 2025

TU Graz Develops AI to Master Molecular Engineering

Researchers at TU Graz are developing an AI-driven system to autonomously and precisely arrange molecules on material surfaces, revolutionizing the construction of complex nanostructures. This approach, leveraging scanning tunneling microscopes and self-learning algorithms, aims to enable breakthroughs in molecular-scale logic circuits and quantum technologies.

16 Jan 2025

How NLP Models Like ChatGPT Are Changing Lives

Research explores how Natural Language Processing (NLP) models like ChatGPT revolutionize understanding and generating human language. It delves into their mechanics, training processes, potential applications, and ethical considerations in AI's rapid evolution.

15 Jan 2025

Researchers Use AI to Craft Drug-Responsive Proteins

Researchers have used a groundbreaking geometric deep learning method, MaSIF, to design proteins that bind specifically to drug-induced surfaces, paving the way for innovative therapies in synthetic biology and disease treatment.

15 Jan 2025

Harnessing AI to Decode and Combat Harmful Algal Blooms

Researchers developed an explainable AI model to predict harmful algal blooms, uncovering key environmental drivers and offering scalable solutions for data-scarce regions.

15 Jan 2025