Francis Halzen, the University of Wisconsin–Madison physicist who won this year’s Nobel Prize in Physics, says he is proud of a 1991 paper proposing that artificial intelligence be used to analyze data from experimental physics. In his account, neural networks and machine-learning techniques later became the tools that let IceCube researchers pick out the Milky Way in neutrino data. The main source is an AFP report carried by Phys.org on October 7, 2026, based on comments Halzen made to reporters in Turin. The report is the only source for his account, so the claims below are his, and they are presented that way.
What Halzen said about the 1991 proposal
Asked about his work, Halzen pointed to an early idea rather than the prize itself. In the words the AFP report quotes: “In fact, the first neural nets appeared in the late 1980s, and I am very proud that I wrote a paper in 1991 proposing to use AI to analyze the data of part of the classical physics experiment.”
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Basic Physics: A Self-Teaching Guide, 3rd Edition | $10.48 | Buy on Amazon |
| 2 |
|
Fundamentals of Physics I: Mechanics, Relativity, and Thermodynamics (Open Yale Courses) | $35.00 | Buy on Amazon |
| 3 |
|
Must Know High School Physics | $14.57 | Buy on Amazon |
| 4 |
|
Physics | $41.26 | Buy on Amazon |
| 5 |
|
Physics for Scientists and Engineers With Modern Physics | $91.26 | Buy on Amazon |
As an Amazon Associate I earn from qualifying purchases.
The report does not give the title, venue or authors of that 1991 paper, and it does not describe what the proposed method looked like in practice. Readers who want to trace the original proposal will need a bibliographic source that the report does not supply.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallWhat IceCube is
IceCube is an observatory whose sensors are 5,484 optical modules set deep in Antarctic ice. The AFP report uses that detector count, and it is the only detector figure in the coverage this article draws on. The observatory is the setting for the Milky Way result Halzen describes below.
#1 Best Overall
How neural networks entered the analysis
Occasional use at first
Halzen describes early neural-network use as occasional. “We kind of used neural nets occasionally,” he said, according to the report. The change, in his telling, came when much more capable networks became available. “And that changed a few years ago, when these very powerful neural nets came along.” The report does not say which networks, which years, or which team made the switch, so the timeline is his general description rather than a dated project history.
Why the Milky Way is hard to see in neutrinos
Halzen explained the problem in terms of the sky itself. “When you look at the sky normally, you see the Milky Way. But when you look at the sky of neutrinos, you see other galaxies, you don’t see the Milky Way.” In other words, the galaxy that dominates an ordinary night sky is not the picture that neutrino data first shows. Seeing it requires extracting a faint signal from a background that is hard to separate.
The Milky Way result
Halzen credits the machine-learning methods with making that extraction possible: “It was only after we used neural nets and machine learning techniques that we finally began to see the Milky Way in our data, which we now have extracted convincingly.” This is his account of the outcome. The report does not describe the model, the training data, how the result was validated, or any measure of accuracy, and it does not say that machine learning alone produced the result. The statement should be read as a researcher’s description of which tools made the difference, not as an independent evaluation of the method.
Free tools Windows power users keep installed
One-click scans. No signup required.
What the report does and does not establish
The table separates what the coverage states, who states it, and what remains unspecified.
Rank #3
| Claim | Source as reported | Status in the coverage | Not stated |
|---|---|---|---|
| A 1991 paper proposed using AI to analyze experimental physics data | Halzen, via AFP/Phys.org (October 7, 2026) | Halzen’s statement; he says he is proud of it | Paper title, venue, authors, method details |
| Neural networks were used occasionally before more powerful networks arrived | Halzen, via AFP/Phys.org | General description; no dates or projects named | Network type, years, team |
| Neural networks and machine learning helped extract the Milky Way in IceCube data | Halzen, via AFP/Phys.org | Halzen’s account of the outcome | Model details, training data, validation, accuracy figure |
| IceCube uses 5,484 optical modules in Antarctic ice | AFP report, via Phys.org | Detector count as reported | Detector upgrade history |
| Project received about $250 million (224 million euros) from the U.S. National Science Foundation | AFP report, via Phys.org | Figure is approximate and attributed to the report | Funding period, primary NSF award citation |
The funding figure is an approximate total reported by AFP. Because the report gives no funding period and no NSF award number, it should not be read as a single annual budget or as a figure tied to the Milky Way analysis specifically.
A different neural-network Nobel
Readers may associate neural networks with a different Nobel. Geoffrey Hinton and John Hopfield shared the 2024 Nobel Prize in Physics for discoveries and inventions that enable machine learning with artificial neural networks, according to University of Toronto coverage of that award. The same coverage quotes Nobel physics committee chair Ellen Moons describing neural networks as useful for sorting and interpreting large amounts of data.
Rank #4
That 2024 recognition concerns a different body of work. It does not verify Halzen’s 1991 proposal or the IceCube analysis he describes, and the two stories should be kept apart.
What the announcement means for fundamental research
Antonio Zoccoli, president of Italy’s National Institute for Nuclear Physics, called the Nobel announcement a “clear recognition of the importance of fundamental research” for “understanding our nature and our origins.” That is Zoccoli’s reading of the award, and the coverage presents it as his view rather than a finding of the work itself.
Best Value
A Nobel Prize Outreach listing also records a first-reaction interview with Halzen, published October 6, 2026. According to its page summary, he recalled the detector’s construction and spoke about the future promise of neutrino astronomy. The listing is useful context for his later comments, but it does not substantiate the 1991 paper or the Milky Way claim.
Asked whether he had always dreamed of a Nobel Prize, Halzen turned to cycling. “I’m from Belgium, I wanted to win the Tour de France,” he joked.
Taken on Halzen’s own terms, the story is about a long gap between an early idea and a later result. A method proposed in 1991 was, by his account, used only occasionally until far more capable networks changed what was possible, and only then did the Milky Way appear in IceCube’s data. The report does not show how much of that delay came from computing power, data quality or other factors, so the lesson is best read as his interpretation of his own project.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Quick Recap
“
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
