Abstract--As data-driven methods, artificial intelligence (AI), and automated workflows accelerate scientific tasks, we see the rate of discovery increasingly limited by human decision-making tasks such as setting objectives, generating hypotheses, and designing experiments. We postulate that cooperative agents are needed to augment the role of humans and enable autonomous discovery . Realizing such agents will require progress in both AI and infrastructure. This situation is emblematic of broader transformations associated with the fourth and fifth paradigms of science, which capture the shift towards data-intensive methods and artificial intelligence, respectively, as integral aspects of scientific exploration [1], [2]. Fields ranging from astrophysics to social sciences now rely on vast datasets, AI models, and computational methods to drive innovation. Hence we face the challenge of not just managing data and building models, but also building systems that enable researchers to integrate and utilize data and models at scale. Current approaches to integrating data-intensive workflows and AI methods have yielded successes, but use techniques that result in siloed solutions that fail to scale or generalize. This paradigm shift demands more than building increasingly sophisticated tools; it calls for a fundamental rethinking of how science is conducted.

Large language models (LLMs) have demonstrated immense capabilities in understanding textual data and are increasingly being adopted to help researchers accelerate scientific discovery through knowledge extraction (information retrieval), knowledge distillation (summarizing key findings and methodologies into concise forms), and knowledge synthesis (aggregating information from multiple scientific sources to address complex queries, generate hypothesis and formulate experimental plans). However, scientific data often exists in both visual and textual modalities. Vision language models (VLMs) address this by incorporating a pretrained vision backbone for processing images and a cross-modal projector that adapts image tokens into the LLM dimensional space, thereby providing richer multimodal comprehension. Nevertheless, off-the-shelf VLMs show limited capabilities in handling domain-specific data and are prone to hallucinations. We developed intelligent assistants finetuned from LLaVA models to enhance multimodal understanding in low-dose radiation therapy (LDRT)-a benign approach used in the treatment of cancer-related illnesses. Using multilingual data from 42,673 articles, we devise complex reasoning and detailed description tasks for visual question answering (VQA) benchmarks. Our assistants, trained on 50,882 image-text pairs, demonstrate superior performance over base models as evaluated using LLM-as-a-judge approach, particularly in reducing hallucination and improving domain-specific comprehension.

The Dynamic PicoProbe at Argonne National Laboratory is undergoing upgrades that will enable it to produce up to 100s of GB of data per day. While this data is highly important for both fundamental science and industrial applications, there is currently limited on-site infrastructure to handle these high-volume data streams. We address this problem by providing a software architecture capable of supporting large-scale data transfers to the neighboring supercomputers at the Argonne Leadership Computing Facility. To prepare for future scientific workflows, we implement two instructive use cases for hyperspectral and spatiotemporal datasets, which include: (i) off-site data transfer, (ii) machine learning/artificial intelligence and traditional data analysis approaches, and (iii) automatic metadata extraction and cataloging of experimental results. This infrastructure supports expected workloads and also provides domain scientists the ability to reinterrogate data from past experiments to yield additional scientific value and derive new insights.

Autonomous intelligence, artificial intelligence (AI) that can act without human intervention, can help identify critical infrastructure cyberattack patterns and network activity, and detect malware to enable enhanced decision-making about defensive responses. That's according to the preliminary findings of an international experiment of AI's ability to secure and defend systems, power grids and other critical assets by cyber experts at the North Atlantic Treaty Organization's (NATO) Cyber Coalition 2022 event late last year. The simulated experiment saw six teams of cyber defenders from NATO allies tasked with setting up computer-based systems and power grids at an imaginary military base and keeping them running during a cyberattack. If hackers interfered with system operations or the power went down for more than 10 minutes, critical systems could go offline. The differentiator was that three of the teams had access to a novel Autonomous Intelligence Cyberdefense Agent (AICA) prototype developed by the US Department of Energy's (DOE) Argonne National Laboratory, while the other three teams did not.

Technology developed at Argonne can help narrow the field of candidates for molten salts, a new study demonstrates. Scientists are searching for new materials to advance the next generation of nuclear power plants. In a recent study, researchers at the U.S. Department of Energy's (DOE) Argonne National Laboratory showed how artificial intelligence could help pinpoint the right types of molten salts, a key component for advanced nuclear reactors. The ability to absorb and store heat makes molten salt important to clean energy and national climate goals. Molten salts can serve as both coolant and fuel in nuclear power reactors that generate electricity without emitting greenhouse gases.

Artificial intelligence is on a tear. Machines can speak, write, play games, and generate original images, video, and music. But as AI's capabilities have grown, so too have its algorithms. A decade ago, machine learning algorithms relied on tens of millions of internal connections, or parameters. Today's algorithms regularly reach into the hundreds of billions and even trillions of parameters.

Aspiring bakers are frequently called upon to adapt award-winning recipes based on differing kitchen setups. Someone might use an eggbeater instead of a stand mixer to make prize-winning chocolate chip cookies, for instance. Being able to reproduce a recipe in different situations and with varying setups is critical for both talented chefs and computational scientists, the latter of whom are faced with a similar problem of adapting and reproducing their own "recipes" when trying to validate and work with new AI models. These models have applications in scientific fields ranging from climate analysis to brain research. "When we talk about data, we have a practical understanding of the digital assets we deal with," said Eliu Huerta, scientist and lead for Translational AI at the U.S. Department of Energy's (DOE) Argonne National Laboratory.

It will be doing so for billions more years. Yet, we have only just begun tapping into that abundant, renewable source of energy at affordable cost. Solar absorbers are a material used to convert this energy into heat or electricity. Maria Chan, a scientist in the U.S. Department of Energy's (DOE) Argonne National Laboratory, has developed a machine learning method for screening many thousands of compounds as solar absorbers. Her co-author on this project was Arun Mannodi-Kanakkithodi, a former Argonne postdoc who is now an assistant professor at Purdue University.

ACM A.M. Turing Award recipient Jack Dongarra never intended to work with computers. Initially, the Distinguished Professor at the University of Tennessee and founder of the Innovative Computing Laboratory (ICL) thought he would be a high school science teacher. A chance internship at the Argonne National Laboratory kindled a lifelong interest in numerical methods and software--and, in particular, in linear algebra, which powered the development of Dongarra's groundbreaking techniques for optimizing operations on increasingly complex computer architectures. Your career in computing began serendipitously, with a semester-long internship at Argonne National Laboratory. As an undergraduate, I worked on EISPACK, a software package designed to solve eigenvalue problems.

Neil Savage is a science and technology writer based in Lowell, MA, USA.