{"id":347,"date":"2026-06-10T08:21:47","date_gmt":"2026-06-10T15:21:47","guid":{"rendered":"https:\/\/scienceblog.com\/neuroedge\/?p=347"},"modified":"2026-06-10T08:21:47","modified_gmt":"2026-06-10T15:21:47","slug":"to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned","status":"publish","type":"post","link":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/","title":{"rendered":"To Find New Physics, an AI First Has to Forget the Old Physics It Learned"},"content":{"rendered":"

Teach a machine everything we know about the universe, and you might expect it to spot the cracks in that knowledge faster. The opposite can happen. A neural network trained on the standard picture of the cosmos sometimes turns that education into a blind spot, quietly forcing strange new signals into familiar old boxes. Which is awkward, because spotting strange new signals is the entire point.<\/p>\n

That is the unexpected catch in a study just published in the Journal of Cosmology and Astroparticle Physics. The headline result is genuinely useful, even cheering for anyone who has watched a supercomputer queue crawl. But it comes stapled to a warning about what these systems do when their prior knowledge gets in the way.<\/p>\n

Start with the problem the researchers were actually trying to solve. Cosmologists describe the universe with a model called \u039bCDM, and it works remarkably well, from the way space expands to how galaxies scatter across it. Trouble is, almost everyone suspects it is incomplete. Massive neutrinos, modified gravity, evolving dark energy: any of these could be lurking just past the edge of the standard model, and testing for them means running vast suites of simulated universes, each one expensive, each one nudged by slightly different physics. The computational bill is enormous.<\/p>\n

A Shortcut Borrowed from Language Models<\/h2>\n

So Veena Krishnaraj, an undergraduate at Princeton University, and her colleagues tried a shortcut. The technique is called transfer learning, and if you have used a modern chatbot you have already met its logic.<\/p>\n

“It’s basically a shortcut,” says Adrian Bayer, a cosmologist at the Flatiron Institute and Princeton University and a co-author of the paper. “Usually people train the AI directly on the most computationally expensive simulations. What we do instead is first use simpler and less expensive \u039bCDM simulations to give the AI an idea of what’s happening, and only afterward move to the more complex models.”<\/p>\n

Bayer reaches for a homely comparison. You first read a basic book to get an idea of the knowledge, he says, and then move to the really complicated book. Krishnaraj puts it another way: the approach spares the network from having to “digest everything at once.” And on paper it pays off handsomely. In some cases, pretraining on the cheap standard-model universes cut the number of expensive beyond-\u039bCDM simulations needed by more than a factor of ten. Since those plain \u039bCDM simulations are already lying around from other survey work, the pretraining stage is, for practical purposes, close to free.<\/p>\n

When Knowing Too Much Backfires<\/h2>\n

Then the textbook analogy turns on them. Imagine reading that introductory medical text, then meeting a rare disease whose symptoms mimic a common one. Your hard-won grounding now tugs you toward the wrong diagnosis. Networks, it turns out, can fall into the same trap, and the researchers gave it a name: negative transfer.<\/p>\n

It showed up most sharply with massive neutrinos. Neutrino mass leaves a fingerprint on cosmic structure that looks a lot like the effect of a standard \u039bCDM parameter called \u03c38, which measures how strongly matter clumps together. The pretrained network, having already learned to read certain small-scale patterns as \u03c38, kept reading them that way, even when the real culprit was the neutrinos. It had to unlearn its own training and reassign those features, and that unlearning cost it dearly.<\/p>\n

“The negative transfer is not random. It is driven by underlying physical degeneracies in the model,” says Krishnaraj. Two different bits of physics produce nearly the same observable smudge, and the AI, sensibly enough, picks the explanation it already knows. The team confirmed the diagnosis by stripping out the small scales where neutrinos and \u03c38 masquerade as each other; do that, and the confusion melts away, which is about as clean a smoking gun as cosmology tends to offer.<\/p>\n

Not every alternative universe gave them grief. Modified gravity behaved much like the neutrino case, with solid gains. Equilateral-type primordial non-Gaussianity, where the physical signatures stay distinct and the degeneracies are mild, transferred nicely. The architecture mattered too: the winning design tucked extra “dummy” nodes into the network during pretraining, leaving slack capacity for the new physics to colonize later, rather than freezing the old knowledge rigidly in place.<\/p>\n

The bigger resonance here is with the foundation models behind today’s generative AI, the sprawling systems pretrained on one giant corpus and fine-tuned for everything else. Cosmology is now asking whether \u039bCDM can play that foundational role for the physics of the universe. The answer, on this evidence, is a qualified yes, with the same caveat the paper states plainly: pretraining can accelerate inference, but may also hinder learning new physics. “So this is something we need to be aware of and try to mitigate,” Krishnaraj says.<\/p>\n

For now the method has only been let loose on simulations, not real sky. But the next generation of cosmological surveys will pour out data at a rate no one can simulate their way through by brute force, and a tool that learns new physics cheaply, if handled with care, is exactly what that flood will demand. The trick, it seems, is building machines clever enough to know when to trust what they have learned, and when to set it aside.<\/p>\n

Krishnaraj, V. et al. (2026), Journal of Cosmology and Astroparticle Physics, JCAP06(2026)026<\/a><\/p>\n

Frequently Asked Questions<\/h2>\n

Why would teaching an AI more about the universe make it worse at finding new physics?<\/strong><\/p>\n

Because the knowledge it absorbs during training can act like a set of fixed expectations. When a genuinely new effect happens to resemble something the AI already recognizes, it tends to file the newcomer under the familiar label rather than flagging it as novel. Cosmologists call this negative transfer, and it shows up precisely when two different bits of physics leave nearly identical fingerprints in the data.<\/p>\n

How does transfer learning actually save on computing costs?<\/strong><\/p>\n

Simulating universes with exotic physics is far more expensive than simulating the standard model. By first training a network on cheap standard-model simulations and then fine-tuning it on a much smaller batch of the costly ones, researchers cut the number of expensive runs needed, in some cases by more than tenfold. The standard-model simulations are often already available from other projects, so that first stage is essentially free.<\/p>\n

Is this the same idea as the AI behind chatbots?<\/strong><\/p>\n

Conceptually, yes. Large language models are pretrained on one enormous body of data and then adapted to specific tasks, and this study tests whether the standard model of cosmology can serve as a similar foundation for the physics of the universe. The promise and the pitfalls turn out to mirror each other: a strong starting point speeds things up, but can also bias the system against recognizing something truly unfamiliar.<\/p>\n

What’s stopping this from being used on real telescope data right now?<\/strong><\/p>\n

So far the approach has only been tested on simulated universes, which gives researchers a controlled setting to study where it works and where it stumbles. Applying it to real observations is the next step, and it is likely to matter a great deal as upcoming surveys produce far more high-precision data than anyone can simulate by force alone.<\/p>\n","protected":false},"excerpt":{"rendered":"

Teach a machine everything we know about the universe, and you might expect it to spot the cracks in that knowledge faster. The opposite can happen. A neural network trained on the standard picture of the cosmos sometimes turns that education into a blind spot, quietly forcing strange new signals into familiar old boxes. Which … Read more<\/a><\/p>\n","protected":false},"author":1297,"featured_media":348,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[4,8,6],"tags":[],"class_list":["post-347","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-computational-innovation","category-physics","category-technology","generate-columns","tablet-grid-50","mobile-grid-100","grid-parent","grid-50"],"yoast_head":"\nTo Find New Physics, an AI First Has to Forget the Old Physics It Learned - NeuroEdge<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"To Find New Physics, an AI First Has to Forget the Old Physics It Learned\" \/>\n<meta property=\"og:description\" content=\"Teach a machine everything we know about the universe, and you might expect it to spot the cracks in that knowledge faster. The opposite can happen. A neural network trained on the standard picture of the cosmos sometimes turns that education into a blind spot, quietly forcing strange new signals into familiar old boxes. Which ... Read more\" \/>\n<meta property=\"og:url\" content=\"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/\" \/>\n<meta property=\"og:site_name\" content=\"NeuroEdge\" \/>\n<meta property=\"article:published_time\" content=\"2026-06-10T15:21:47+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/06\/Quijote-simulations-.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"700\" \/>\n\t<meta property=\"og:image:height\" content=\"467\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"NeuroEdge\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"NeuroEdge\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"6 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/\"},\"author\":{\"name\":\"NeuroEdge\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#\\\/schema\\\/person\\\/a13c664778e7eb97cb71e3e1ad356d2e\"},\"headline\":\"To Find New Physics, an AI First Has to Forget the Old Physics It Learned\",\"datePublished\":\"2026-06-10T15:21:47+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/\"},\"wordCount\":1213,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2026\\\/06\\\/Quijote-simulations-.jpg\",\"articleSection\":[\"Computational Innovation\",\"Physics\",\"Technology\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/#respond\"]}],\"copyrightYear\":\"2026\",\"copyrightHolder\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/\",\"name\":\"To Find New Physics, an AI First Has to Forget the Old Physics It Learned - NeuroEdge\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2026\\\/06\\\/Quijote-simulations-.jpg\",\"datePublished\":\"2026-06-10T15:21:47+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/#primaryimage\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2026\\\/06\\\/Quijote-simulations-.jpg\",\"contentUrl\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2026\\\/06\\\/Quijote-simulations-.jpg\",\"width\":700,\"height\":467,\"caption\":\"Two images from the Quijote simulations used in this study. The panels show the same region of the Universe, but in different cosmological models. The top image corresponds to the standard \u039bCDM model, while the bottom image shows a universe with massive neutrinos and modified gravity. The differences are subtle, but they reveal how changes in the underlying physics can affect the formation and distribution of cosmic structures.\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2026\\\/06\\\/10\\\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"To Find New Physics, an AI First Has to Forget the Old Physics It Learned\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#website\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/\",\"name\":\"NeuroEdge\",\"description\":\"A data-driven look at neuroscience and AI, for investors, policymakers, and innovators.\",\"publisher\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#organization\",\"name\":\"NeuroEdge\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2025\\\/04\\\/cropped-neuroedge_logo.jpg\",\"contentUrl\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2025\\\/04\\\/cropped-neuroedge_logo.jpg\",\"width\":955,\"height\":191,\"caption\":\"NeuroEdge\"},\"image\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#\\\/schema\\\/logo\\\/image\\\/\"}},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#\\\/schema\\\/person\\\/a13c664778e7eb97cb71e3e1ad356d2e\",\"name\":\"NeuroEdge\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g\",\"caption\":\"NeuroEdge\"},\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/author\\\/neuroedge\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"To Find New Physics, an AI First Has to Forget the Old Physics It Learned - NeuroEdge","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/","og_locale":"en_US","og_type":"article","og_title":"To Find New Physics, an AI First Has to Forget the Old Physics It Learned","og_description":"Teach a machine everything we know about the universe, and you might expect it to spot the cracks in that knowledge faster. The opposite can happen. A neural network trained on the standard picture of the cosmos sometimes turns that education into a blind spot, quietly forcing strange new signals into familiar old boxes. Which ... Read more","og_url":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/","og_site_name":"NeuroEdge","article_published_time":"2026-06-10T15:21:47+00:00","og_image":[{"width":700,"height":467,"url":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/06\/Quijote-simulations-.jpg","type":"image\/jpeg"}],"author":"NeuroEdge","twitter_card":"summary_large_image","twitter_misc":{"Written by":"NeuroEdge","Est. reading time":"6 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/#article","isPartOf":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/"},"author":{"name":"NeuroEdge","@id":"https:\/\/scienceblog.com\/neuroedge\/#\/schema\/person\/a13c664778e7eb97cb71e3e1ad356d2e"},"headline":"To Find New Physics, an AI First Has to Forget the Old Physics It Learned","datePublished":"2026-06-10T15:21:47+00:00","mainEntityOfPage":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/"},"wordCount":1213,"commentCount":0,"publisher":{"@id":"https:\/\/scienceblog.com\/neuroedge\/#organization"},"image":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/#primaryimage"},"thumbnailUrl":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/06\/Quijote-simulations-.jpg","articleSection":["Computational Innovation","Physics","Technology"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/#respond"]}],"copyrightYear":"2026","copyrightHolder":{"@id":"https:\/\/scienceblog.com\/#organization"}},{"@type":"WebPage","@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/","url":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/","name":"To Find New Physics, an AI First Has to Forget the Old Physics It Learned - NeuroEdge","isPartOf":{"@id":"https:\/\/scienceblog.com\/neuroedge\/#website"},"primaryImageOfPage":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/#primaryimage"},"image":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/#primaryimage"},"thumbnailUrl":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/06\/Quijote-simulations-.jpg","datePublished":"2026-06-10T15:21:47+00:00","breadcrumb":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/#primaryimage","url":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/06\/Quijote-simulations-.jpg","contentUrl":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/06\/Quijote-simulations-.jpg","width":700,"height":467,"caption":"Two images from the Quijote simulations used in this study. The panels show the same region of the Universe, but in different cosmological models. The top image corresponds to the standard \u039bCDM model, while the bottom image shows a universe with massive neutrinos and modified gravity. The differences are subtle, but they reveal how changes in the underlying physics can affect the formation and distribution of cosmic structures."},{"@type":"BreadcrumbList","@id":"https:\/\/scienceblog.com\/neuroedge\/2026\/06\/10\/to-find-new-physics-an-ai-first-has-to-forget-the-old-physics-it-learned\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/scienceblog.com\/neuroedge\/"},{"@type":"ListItem","position":2,"name":"To Find New Physics, an AI First Has to Forget the Old Physics It Learned"}]},{"@type":"WebSite","@id":"https:\/\/scienceblog.com\/neuroedge\/#website","url":"https:\/\/scienceblog.com\/neuroedge\/","name":"NeuroEdge","description":"A data-driven look at neuroscience and AI, for investors, policymakers, and innovators.","publisher":{"@id":"https:\/\/scienceblog.com\/neuroedge\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/scienceblog.com\/neuroedge\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/scienceblog.com\/neuroedge\/#organization","name":"NeuroEdge","url":"https:\/\/scienceblog.com\/neuroedge\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/scienceblog.com\/neuroedge\/#\/schema\/logo\/image\/","url":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/cropped-neuroedge_logo.jpg","contentUrl":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/cropped-neuroedge_logo.jpg","width":955,"height":191,"caption":"NeuroEdge"},"image":{"@id":"https:\/\/scienceblog.com\/neuroedge\/#\/schema\/logo\/image\/"}},{"@type":"Person","@id":"https:\/\/scienceblog.com\/neuroedge\/#\/schema\/person\/a13c664778e7eb97cb71e3e1ad356d2e","name":"NeuroEdge","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g","caption":"NeuroEdge"},"url":"https:\/\/scienceblog.com\/neuroedge\/author\/neuroedge\/"}]}},"jetpack_featured_media_url":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/06\/Quijote-simulations-.jpg","jetpack_likes_enabled":true,"jetpack_sharing_enabled":true,"jetpack-related-posts":[{"id":122,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/","url_meta":{"origin":347,"position":0},"title":"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You","author":"NeuroEdge","date":"May 5, 2025","format":false,"excerpt":"Researchers at George Washington University have developed a groundbreaking mathematical formula that predicts exactly when artificial intelligence systems like ChatGPT will suddenly shift from helpful to harmful responses \u2013 a phenomenon they've dubbed the \"Jekyll-and-Hyde tipping point.\" The new research may finally answer why AI sometimes abruptly goes off the\u2026","rel":"","context":"In "Physics"","block_context":{"text":"Physics","link":"https:\/\/scienceblog.com\/neuroedge\/category\/physics\/"},"img":{"alt_text":"HAL from 2001","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp?resize=525%2C300&ssl=1 1.5x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp?resize=700%2C400&ssl=1 2x"},"classes":[]},{"id":248,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/10\/08\/ai-learns-to-spot-exploding-stars-from-just-15-examples\/","url_meta":{"origin":347,"position":1},"title":"AI Learns to Spot Exploding Stars From Just 15 Examples","author":"NeuroEdge","date":"October 8, 2025","format":false,"excerpt":"Modern telescopes are magnificent gossips, generating millions of alerts every night about potential changes in the cosmos. The problem? Most of these whispers are lies - satellite trails, cosmic ray hits, instrumental hiccups masquerading as genuine discoveries. For years, astronomers have deployed specialized neural networks to separate wheat from chaff,\u2026","rel":"","context":"In "Automation & Efficiency"","block_context":{"text":"Automation & Efficiency","link":"https:\/\/scienceblog.com\/neuroedge\/category\/automation-efficiency\/"},"img":{"alt_text":"The same transient is shown in three surveys, with rows corresponding to Pan-STARRS (top), MeerLICHT (middle), and ATLAS (bottom). Each row presents, from left to right, the New, Reference, and Difference images. Red circles mark the expected position of the transient candidate at the centre of each stamp. All stamps are 100\u00d7100 pixels, but their angular sky coverage differs due to survey-specific pixel scales: Pan-STARRS 0.25\u2033\/pixel, MeerLICHT 0.56\u2033\/pixel, and ATLAS 1.86\u2033\/pixel. Credit: Stoppa & Bulmus et al., Nature Astronomy (2025).","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/10\/how-gemini-operates.jpg?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/10\/how-gemini-operates.jpg?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/10\/how-gemini-operates.jpg?resize=525%2C300&ssl=1 1.5x"},"classes":[]},{"id":22,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/04\/04\/ai-slashes-fluid-simulation-times-fifteenfold\/","url_meta":{"origin":347,"position":2},"title":"AI Slashes Fluid Simulation Times Fifteenfold","author":"NeuroEdge","date":"April 4, 2025","format":false,"excerpt":"Osaka researchers have developed an AI model that performs complex fluid simulations in minutes instead of hours, potentially transforming offshore engineering while maintaining high accuracy. This advancement could accelerate development cycles for maritime technologies and enable real-time monitoring systems previously considered computationally impossible. Traditional particle-based fluid simulations, essential for predicting\u2026","rel":"","context":"In "Computational Innovation"","block_context":{"text":"Computational Innovation","link":"https:\/\/scienceblog.com\/neuroedge\/category\/computational-innovation\/"},"img":{"alt_text":"Simulated waves","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/ocen-simulation.png?resize=350%2C200&ssl=1","width":350,"height":200},"classes":[]},{"id":263,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/11\/19\/he-taught-ai-to-say-i-dont-know\/","url_meta":{"origin":347,"position":3},"title":"He Taught AI to Say “I Don’t Know”","author":"NeuroEdge","date":"November 19, 2025","format":false,"excerpt":"Artificial intelligence can diagnose disease, write essays, and generate art. But it often refuses to admit when it's wrong. Now, a University of Arizona astronomer has found a way to change that. In a preprint posted to arXiv, Peter Behroozi introduces a new method for reducing hallucinations in large-scale AI\u2026","rel":"","context":"In "Computational Innovation"","block_context":{"text":"Computational Innovation","link":"https:\/\/scienceblog.com\/neuroedge\/category\/computational-innovation\/"},"img":{"alt_text":"Light rays are propagating smoothly through a noisy, high-dimensional space in this artist\u2019s impression. The new ray tracing algorithm improves on previous methods by better averaging information over trajectories, making it many orders of magnitude faster for quantifying uncertainties in large neural networks.","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/11\/AI-Behroozi-feature-image.png.webp?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/11\/AI-Behroozi-feature-image.png.webp?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/11\/AI-Behroozi-feature-image.png.webp?resize=525%2C300&ssl=1 1.5x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/11\/AI-Behroozi-feature-image.png.webp?resize=700%2C400&ssl=1 2x"},"classes":[]},{"id":336,"url":"https:\/\/scienceblog.com\/neuroedge\/2026\/05\/28\/human-brain-organoids-are-learning-to-play-video-games-that-is-only-the-beginning\/","url_meta":{"origin":347,"position":4},"title":"Human Brain Organoids Are Learning to Play Video Games. That Is Only the Beginning.","author":"NeuroEdge","date":"May 28, 2026","format":false,"excerpt":"Start with a skin cell. Bathe it in the right cocktail of chemical signals and it forgets what it is, regressing to a state of biological openness, capable of becoming almost anything. Over the next several months, coax it toward a neural fate. What you end up with, roughly a\u2026","rel":"","context":"In "Brain Health"","block_context":{"text":"Brain Health","link":"https:\/\/scienceblog.com\/neuroedge\/category\/brain-health\/"},"img":{"alt_text":"Brain organoid","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/05\/brain-organoid.jpg?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/05\/brain-organoid.jpg?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/05\/brain-organoid.jpg?resize=525%2C300&ssl=1 1.5x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/05\/brain-organoid.jpg?resize=700%2C400&ssl=1 2x"},"classes":[]},{"id":94,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/04\/28\/quantum-neural-hybrid-solves-impossible-math\/","url_meta":{"origin":347,"position":5},"title":"Quantum-Neural Hybrid Solves Impossible Math","author":"NeuroEdge","date":"April 28, 2025","format":false,"excerpt":"The worlds of quantum mechanics and neural networks have collided in a new system that's setting benchmarks for solving previously intractable optimization problems. A multi-university team led by Shantanu Chakrabartty at Washington University in St. Louis has introduced NeuroSA, a neuromorphic architecture that leverages quantum tunneling mechanisms to reliably discover\u2026","rel":"","context":"In "Physics"","block_context":{"text":"Physics","link":"https:\/\/scienceblog.com\/neuroedge\/category\/physics\/"},"img":{"alt_text":"NeuroSA is a neuromorphic architecture with quantum capabilities that allow it to solve optimization problems more reliably than state-of-the-art methods. (Image created by Shantanu Chakrabartty using Google Gemini)","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/gemini-neuro-quantum.jpg?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/gemini-neuro-quantum.jpg?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/gemini-neuro-quantum.jpg?resize=525%2C300&ssl=1 1.5x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/gemini-neuro-quantum.jpg?resize=700%2C400&ssl=1 2x"},"classes":[]}],"_links":{"self":[{"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/posts\/347","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/users\/1297"}],"replies":[{"embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/comments?post=347"}],"version-history":[{"count":1,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/posts\/347\/revisions"}],"predecessor-version":[{"id":349,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/posts\/347\/revisions\/349"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/media\/348"}],"wp:attachment":[{"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/media?parent=347"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/categories?post=347"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/tags?post=347"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}