Home › Bambini Prodigio › Jagdeep Bhatia
PREMIO SCIENTIFICO

🇺🇸 Jagdeep Bhatia

Ha vinto 175.000 $ per aver dimostrato che una macchina che apprende dovrebbe scegliere le proprie domande.

Gli algoritmi di Jagdeep Bhatia per l'apprendimento automatico interattivo hanno conquistato il secondo posto al Regeneron Science Talent Search del 2020.

Jagdeep Bhatia, di Green Brook, New Jersey, e studente alla Watchung Hills Regional High School di Warren, ha ottenuto il secondo posto e 175.000 $ al Regeneron Science Talent Search 2020. Il suo progetto ha stabilito algoritmi semplici e veloci per l'apprendimento automatico interattivo con controesempi casuali: metodi che consentono a un programma, mentre apprende un nuovo concetto, di scegliere quali domande porre invece di campionare in modo casuale.

Jagdeep Bhatia — Meet Jagdeep Bhatia, 2nd Place Winner of the 2020 Regeneron Science Talent…

Meet Jagdeep Bhatia, 2nd Place Winner of the 2020 Regeneron Science Talent Search — SocietyforScience

🧠 Take the free 8-minute IQ test
Condividi su X Facebook LinkedIn WhatsApp Reddit

Second place, and $175,000

Society for Science lists Bhatia second in the 2020 Science Talent Search, behind Lillian Petersen and ahead of Brendan Crotty, with an award of $175,000. His school is given as Watchung Hills Regional High School in Warren, New Jersey, his home town as Green Brook in the same state.

The competition that year drew 1,993 entrants, from whom forty finalists were chosen; every finalist received at least $25,000, and the top ten took between $40,000 and $250,000. Second place is therefore not a consolation. It is the second-largest award in what is generally treated as the most demanding pre-university research competition in the United States.

The idea: let the learner choose the question

Bhatia's project is titled Simple and Fast Algorithms for Interactive Machine Learning with Random Counter-Examples. The Society describes it as developing fast machine learning algorithms for computer programs that learn new concepts from an instructor, asking strategically selected questions rather than random ones, with the aim of training robots and automated devices more efficiently.

The distinction matters more than it sounds. In ordinary supervised learning, a model is handed a pile of labelled examples and has no say in which ones it sees. In the interactive setting, the learner can put a question to a teacher, and the choice of question determines how fast it converges. A well-chosen counter-example, one that contradicts the learner's current guess, can be worth thousands of random samples. Algorithms that are both simple and fast at generating those questions are the practical bottleneck between the theory and a robot that can be taught on a factory floor in an afternoon.

Where it sat in the 2020 field

The 2020 top ten was unusually wide in subject. Lillian Petersen took first with $250,000. Brendan Crotty, a home-schooled student from Muskogee, Oklahoma, took third and $150,000 for a hybrid industrial gas burner that emitted 19 per cent less nitrogen dioxide. Rupert Li was fourth, Anaiah Thomas fifth, Katherine St George sixth, Alek Westover seventh, Adriane Thompson eighth, Rohan Wagh ninth and Arjun Neervannan tenth.

Machine learning entries have become a fixture of the search, and the reason is the same one that makes them contentious: the equipment is a laptop. A student with mathematical ability and a public dataset can produce work at the edge of a research literature without a bench, a supervisor or a grant. That widens who can compete, and it raises the bar for what counts as original.

What the prize is for

The Science Talent Search does not judge a demonstration. Entrants submit a full research report along with their school record and essays, and the forty finalists are then interviewed in Washington by panels of working scientists who probe the surrounding field as much as the project itself. A finalist who cannot explain why the problem matters, or what would falsify the result, does not place.

Against that standard, an award of $175,000 is a statement about the student rather than the paper. The Society's own framing treats the search as identifying future research scientists, and the roll of past finalists is heavy with people who became exactly that.

IL TEST GRATUITO
Quanto fa il tuo?

Venti domande, otto minuti di orologio e un percentile misurato su tutti quelli che l’hanno fatto. Senza registrazione.

Fai il test del QI →

What is documented, and what is not

Firm: Bhatia placed second in the 2020 Regeneron Science Talent Search and received $175,000; he attended Watchung Hills Regional High School in Warren, New Jersey, and is listed from Green Brook, New Jersey; his project concerned simple and fast algorithms for interactive machine learning with random counter-examples.

Not documented, in the source used here: his age or birth year, where he went to university, whether the work was published in a journal or conference, and what he has done since March 2020. The Society's finalist listing is precise about the project and silent about the person, which is the reverse of most prize coverage and, for a page like this one, the better half to have.

20202020Mar2020SinceMILESTONES IN ORDERON ONE TIMELINE
Jagdeep Bhatia — every milestone on this page, in order.
2020
Nominato tra i quaranta finalistiNominato tra i quaranta finalisti al Regeneron Science Talent Search, scelto tra 1.993 partecipanti
Mar 2020
Conquista il secondo posto e 175.000 $Conquista il secondo posto e 175.000 $, dietro a Lillian Petersen
2020
La Society for Science pubblica il suo progettoLa Society for Science pubblica il suo progetto sull'apprendimento automatico interattivo con controesempi casuali
Since 2020
Nessun ulteriore record pubblicato trovatoNessun ulteriore record pubblicato trovato nelle fonti utilizzate per questa pagina

Watch: Jagdeep Bhatia on camera

PersonaPaeseTraguardoEtà / Dato
Jagdeep BhatiaStati UnitiSecondo posto, Regeneron Science Talent Search 2020—
Melissa ReyStati UnitiGran premio, Discovery Education 3M Young Scientist Challenge 2008—
Sara VolzStati UnitiPrimo posto, Intel Science Talent Search 2013—
Jonah KallenbachStati UnitiSecondo posto, Intel Science Talent Search 2013—
Victor CaiStati UnitiSecondo posto, Regeneron Science Talent Search 2022—

What is documented — and what is not

✓ Documented

  • Second place and $175,000 at the 2020 Regeneron Science Talent Search
  • Attended Watchung Hills Regional High School, Warren, New Jersey; listed from Green Brook
  • Project title and description published by Society for Science
  • Forty finalists were chosen from 1,993 entrants in 2020

? Unverified, disputed or wrong

  • Age and birth year are not stated
  • No record here of the work being published in a journal or conference
  • Nothing documented about his university or career after March 2020

Questions people ask about Jagdeep Bhatia

What did Jagdeep Bhatia win?

Second place and $175,000 at the 2020 Regeneron Science Talent Search, according to Society for Science.

What was his project about?

Fast algorithms for interactive machine learning: letting a program that is learning a new concept ask strategically chosen questions instead of sampling at random, to train robots and automated devices more efficiently.

Where did he go to school?

Watchung Hills Regional High School in Warren, New Jersey. His home town is listed as Green Brook, New Jersey.

How old is Jagdeep Bhatia?

No source consulted gives his age or birth year. Entrants are high school students.

How competitive is the Science Talent Search?

In 2020 there were 1,993 entrants and forty finalists. Each finalist received at least $25,000; the top ten took between $40,000 and $250,000.

Who won that year?

Lillian Petersen took first place and $250,000. Brendan Crotty was third with $150,000.

Where is Jagdeep Bhatia now?

Nothing published after the March 2020 result was found in the sources used for this page.

Quick quiz: how well do you know Jagdeep’s story?

How much did Jagdeep Bhatia win at the 2020 Science Talent Search?

What did his project let a learning program do?

How many entrants were there in 2020?

Now test yourself: the free 8-minute IQ test →

Sources

  1. 2020 Regeneron Science Talent Search finalists — Society for Science, 2020-03-11

Checked by the Geniuses.club editorial team. Ages are given as of September 2026.

Scopri Altri Bambini Prodigio

Oltre 50 storie di giovani geni contemporanei che trasformano scacchi, musica, scienza e arte.

Esplora Tutti i Prodigi →

Confronta con i grandi

Augustine Of Hippo vs Garry KasparovBobby Fischer vs Magnus CarlsenAda Lovelace vs Stephen HawkingBaruch Spinoza vs Mark Twain
Vedi Classifica →Tutti i confronti →

Bambini prodigio

Jackie EvanchoJackie EvanchoYoungest Solo Platinum-Selling Singer in U.S. History — Sang…Umi GarrettUmi GarrettWent viral playing Liszt on Ellen at eight, then won a Chopin…Maddie ZieglerMaddie ZieglerSia's 'Chandelier' Music Video at Age 11 — Two Billion YouTube…Daniel TammetDaniel TammetRecited pi from memory to 22,514 digits in just over five hours
Bambini prodigio →

Gioca e torna domani

Sfida del genio del giorno · Guess the genius
Naturalist who wrote 'On the Origin of Species' after his voyage on HMS Beagle.
Tocca la tua risposta ↓
Quale genio sei? Test del QI gratuito
Which genius shares your birthday? Get one unforgettable mind in your inbox every week. Free, unsubscribe anytime.
or create a free account →