We evaluate EvoTune on three combinatorial optimization problems:
📦 Bin packing
🗺️ Traveling salesman
🧩 Flatpack
📈 Key result: Across all problems and three LLMs, EvoTune discovers programs with higher rewards faster compared to the Funsearch (no finetune) baseline 7/12
📦 Bin packing
🗺️ Traveling salesman
🧩 Flatpack
📈 Key result: Across all problems and three LLMs, EvoTune discovers programs with higher rewards faster compared to the Funsearch (no finetune) baseline 7/12
April 26, 2025 at 4:56 PM
We evaluate EvoTune on three combinatorial optimization problems:
📦 Bin packing
🗺️ Traveling salesman
🧩 Flatpack
📈 Key result: Across all problems and three LLMs, EvoTune discovers programs with higher rewards faster compared to the Funsearch (no finetune) baseline 7/12
📦 Bin packing
🗺️ Traveling salesman
🧩 Flatpack
📈 Key result: Across all problems and three LLMs, EvoTune discovers programs with higher rewards faster compared to the Funsearch (no finetune) baseline 7/12
FunSearch pairs a pre-trained LLM for generating creative code with an evaluator to refine ideas, evolving initial solutions into new knowledge in math and computer science. #AI #LLMs
FunSearch: Making new discoveries in mathematical sciences using Large Language Models
We introduce FunSearch, a method for searching for “functions” written in computer code, and find new solutions in mathematics and computer science. FunSearch works by pairing a pre-trained LLM, w...
deepmind.google
December 15, 2023 at 3:35 AM
@brouhaha when we start seeing FunSearch style breakthroughs (novel solutions to unsolved problems like cap set) in the arena of algorithms that cross quantum, classical, and GPU, I think we are most likely to see surprises
May 29, 2025 at 5:03 PM
@brouhaha when we start seeing FunSearch style breakthroughs (novel solutions to unsolved problems like cap set) in the arena of algorithms that cross quantum, classical, and GPU, I think we are most likely to see surprises
extremal combinatorics. Our implementation is designed to be useful in practice for working mathematicians; it does not require expertise in machine learning or access to high-performance computing resources. Applying {\it funsearch} to a new problem [2/4 of https://arxiv.org/abs/2503.11061v1]
March 17, 2025 at 5:59 AM
extremal combinatorics. Our implementation is designed to be useful in practice for working mathematicians; it does not require expertise in machine learning or access to high-performance computing resources. Applying {\it funsearch} to a new problem [2/4 of https://arxiv.org/abs/2503.11061v1]
Our approach builds upon FunSearch, a recently proposed evolutionary framework utilizing large language models to propose iterative improvements to model hypotheses.
February 7, 2025 at 6:54 AM
Our approach builds upon FunSearch, a recently proposed evolutionary framework utilizing large language models to propose iterative improvements to model hypotheses.
Oh, interesting! I’ve opened but not read your paper. I was going to ask if you saw this as a variant of “genetic programing” but that phrase is not in your paper. You do mention FunSearch, which indeed talked about its connection to GP. Similar but with GP crossover replaced with LLM synthesis.
April 27, 2025 at 9:12 PM
Oh, interesting! I’ve opened but not read your paper. I was going to ask if you saw this as a variant of “genetic programing” but that phrase is not in your paper. You do mention FunSearch, which indeed talked about its connection to GP. Similar but with GP crossover replaced with LLM synthesis.
大規模言語モデルを駆使して数理科学問題の新しい解決策を数学者よりも巧みに出力するAIシステム「FunSearch」をGoogle DeepMindが発表
https://gigazine.net/news/20231215-google-deepmind-llm-funsearch/
https://gigazine.net/news/20231215-google-deepmind-llm-funsearch/
大規模言語モデルを駆使して数理科学問題の新しい解決策を数学者よりも巧みに出力するAIシステム「FunSearch」をGoogle DeepMindが発表
GoogleのAI開発部門であるGoogle DeepMindが、大規模言語モデル(LLM)を利用して数理科学の未解決の問題を解決したり、新しい知見をもたらしたりすることができるAIシステム「FunSearch」を発表しました。Google DeepMindのデミス・ハサビスCEOは「FunSearchのブレークスルーに興奮しています」と語っています。
gigazine.net
December 15, 2023 at 2:15 AM
大規模言語モデルを駆使して数理科学問題の新しい解決策を数学者よりも巧みに出力するAIシステム「FunSearch」をGoogle DeepMindが発表
https://gigazine.net/news/20231215-google-deepmind-llm-funsearch/
https://gigazine.net/news/20231215-google-deepmind-llm-funsearch/
FunSearch is an optimization process that uses LLMs to evolve Python programs. we adapted it to our use-case to evolve parameterized symbolic cognitive models using a bi-level optimization: each program's params are fit on the training data for evaluation.
4/12
4/12
February 10, 2025 at 12:21 PM
FunSearch is an optimization process that uses LLMs to evolve Python programs. we adapted it to our use-case to evolve parameterized symbolic cognitive models using a bi-level optimization: each program's params are fit on the training data for evaluation.
4/12
4/12
The Mini Sneaky Ninja is hidden somewhere in this scene 👀
It’s not as easy as it looks — think you can spot him?
#SneakyNinja #MiniNinja #PuzzleChallenge #HiddenGame #SeekAndSpot #FunSearch #BlueJays #Baseball #Game
youtube.com/shorts/pTGp4...
It’s not as easy as it looks — think you can spot him?
#SneakyNinja #MiniNinja #PuzzleChallenge #HiddenGame #SeekAndSpot #FunSearch #BlueJays #Baseball #Game
youtube.com/shorts/pTGp4...
Let's Go Blue Jays! #SneakyNinja #challenge #puzzle #HiddenObject #Game #hideandseek #Ninja #findit
YouTube video by Spot The Sneaky Ninja
youtube.com
October 6, 2025 at 12:22 PM
The Mini Sneaky Ninja is hidden somewhere in this scene 👀
It’s not as easy as it looks — think you can spot him?
#SneakyNinja #MiniNinja #PuzzleChallenge #HiddenGame #SeekAndSpot #FunSearch #BlueJays #Baseball #Game
youtube.com/shorts/pTGp4...
It’s not as easy as it looks — think you can spot him?
#SneakyNinja #MiniNinja #PuzzleChallenge #HiddenGame #SeekAndSpot #FunSearch #BlueJays #Baseball #Game
youtube.com/shorts/pTGp4...
arXiv:2503.10721v1 Announce Type: new
Abstract: Large language models (LLMs) have greatly accelerated the automation of algorithm generation and optimization. However, current methods such as EoH and FunSearch mainly rely on predefined templates and [1/7 of https://arxiv.org/abs/2503.10721v1]
Abstract: Large language models (LLMs) have greatly accelerated the automation of algorithm generation and optimization. However, current methods such as EoH and FunSearch mainly rely on predefined templates and [1/7 of https://arxiv.org/abs/2503.10721v1]
March 17, 2025 at 5:59 AM
arXiv:2503.10721v1 Announce Type: new
Abstract: Large language models (LLMs) have greatly accelerated the automation of algorithm generation and optimization. However, current methods such as EoH and FunSearch mainly rely on predefined templates and [1/7 of https://arxiv.org/abs/2503.10721v1]
Abstract: Large language models (LLMs) have greatly accelerated the automation of algorithm generation and optimization. However, current methods such as EoH and FunSearch mainly rely on predefined templates and [1/7 of https://arxiv.org/abs/2503.10721v1]
💡 Summary by GPT3:
研究者たちは、数学やコンピュータ科学の新しい解法を探索するために、FunSearchと呼ばれる方法を開発しました。FunSearchは、事前に訓練された大規模言語モデル（LLM）と自動評価者を組み合わせて、コンピュータコードの形で創造的な解法を生成します。このシステムはLLMと評価者の間で反復し、初期の解法を新たな知識に進化させることができます。FunSearchは、キャップセット問題などの数学の難解な問題に対して新しい解法を成功裏に発見し、また、 (1/2)
研究者たちは、数学やコンピュータ科学の新しい解法を探索するために、FunSearchと呼ばれる方法を開発しました。FunSearchは、事前に訓練された大規模言語モデル（LLM）と自動評価者を組み合わせて、コンピュータコードの形で創造的な解法を生成します。このシステムはLLMと評価者の間で反復し、初期の解法を新たな知識に進化させることができます。FunSearchは、キャップセット問題などの数学の難解な問題に対して新しい解法を成功裏に発見し、また、 (1/2)
December 15, 2023 at 11:43 AM
💡 Summary by GPT3:
研究者たちは、数学やコンピュータ科学の新しい解法を探索するために、FunSearchと呼ばれる方法を開発しました。FunSearchは、事前に訓練された大規模言語モデル（LLM）と自動評価者を組み合わせて、コンピュータコードの形で創造的な解法を生成します。このシステムはLLMと評価者の間で反復し、初期の解法を新たな知識に進化させることができます。FunSearchは、キャップセット問題などの数学の難解な問題に対して新しい解法を成功裏に発見し、また、 (1/2)
研究者たちは、数学やコンピュータ科学の新しい解法を探索するために、FunSearchと呼ばれる方法を開発しました。FunSearchは、事前に訓練された大規模言語モデル（LLM）と自動評価者を組み合わせて、コンピュータコードの形で創造的な解法を生成します。このシステムはLLMと評価者の間で反復し、初期の解法を新たな知識に進化させることができます。FunSearchは、キャップセット問題などの数学の難解な問題に対して新しい解法を成功裏に発見し、また、 (1/2)
At 11:30 I will be presenting our work on using FunSearch (now Alpha evolve) to discover interpretable models of cognition:
3/4
bsky.app/profile/pcas...
3/4
bsky.app/profile/pcas...
Can LLMs be used to discover interpretable models of human and animal behavior?🤔
Turns out: yes!
Thrilled to share our latest preprint where we used FunSearch to automatically discover symbolic cognitive models of behavior.
1/12
Turns out: yes!
Thrilled to share our latest preprint where we used FunSearch to automatically discover symbolic cognitive models of behavior.
1/12
July 14, 2025 at 3:17 PM
At 11:30 I will be presenting our work on using FunSearch (now Alpha evolve) to discover interpretable models of cognition:
3/4
bsky.app/profile/pcas...
3/4
bsky.app/profile/pcas...
AI discovers new Mathematics and Algorithms!
FunSearch, by DeepMind, is a groundbreaking innovation using LLMs that has achieved what was once thought impossible:
Discovering new mathematical knowledge and algorithms.
deepmind.google/discover/blo...
FunSearch, by DeepMind, is a groundbreaking innovation using LLMs that has achieved what was once thought impossible:
Discovering new mathematical knowledge and algorithms.
deepmind.google/discover/blo...
FunSearch: Making new discoveries in mathematical sciences using Large Language Models
We introduce FunSearch, a method for searching for “functions” written in computer code, and find new solutions in mathematics and computer science. FunSearch works by pairing a pre-trained LLM, w...
deepmind.google
December 14, 2023 at 7:48 PM
AI discovers new Mathematics and Algorithms!
FunSearch, by DeepMind, is a groundbreaking innovation using LLMs that has achieved what was once thought impossible:
Discovering new mathematical knowledge and algorithms.
deepmind.google/discover/blo...
FunSearch, by DeepMind, is a groundbreaking innovation using LLMs that has achieved what was once thought impossible:
Discovering new mathematical knowledge and algorithms.
deepmind.google/discover/blo...
I don't think they've released the actual output or discovery in-depth with regards to this model. It's a space I intend to watch.
They have accomplishments for AlphaFold and FunSearch so there is at least a track record and not just "AGI make go gib money"
AlphaFold was released as source iirc
They have accomplishments for AlphaFold and FunSearch so there is at least a track record and not just "AGI make go gib money"
AlphaFold was released as source iirc
May 14, 2025 at 4:23 PM
I don't think they've released the actual output or discovery in-depth with regards to this model. It's a space I intend to watch.
They have accomplishments for AlphaFold and FunSearch so there is at least a track record and not just "AGI make go gib money"
AlphaFold was released as source iirc
They have accomplishments for AlphaFold and FunSearch so there is at least a track record and not just "AGI make go gib money"
AlphaFold was released as source iirc
FunSearch: Making new discoveries in mathematical sciences using LLMs
ファンサーチLLMを使って数理科学に新たな発見をもたらす
🔺 326
💬 16
🔗 HN Post | Article
ファンサーチLLMを使って数理科学に新たな発見をもたらす
🔺 326
💬 16
🔗 HN Post | Article
FunSearch: Making new discoveries in mathematical sciences using Large Language Models
We introduce FunSearch, a method for searching for “functions” written in computer code, and find new solutions in mathematics and computer science. FunSearch works by pairing a pre-trained LLM, whose goal is to provide creative solutions in the form of computer code, with an automated “evaluator”, which guards against hallucinations and incorrect ideas.
deepmind.google
December 15, 2023 at 11:43 AM
Google FunSearch Shows a Path for LLMs to Improve Math Problem Solving Skills https://synthedia.substack.com/p/google-funsearch-shows-a-path-for (more thoughts on FunSearch) #AI #math #discoveries
Google FunSearch Shows a Path for LLMs to Improve Math Problem Solving Skills
Automated code iteration leads to novel ideas
synthedia.substack.com
December 17, 2023 at 3:52 PM
Google FunSearch Shows a Path for LLMs to Improve Math Problem Solving Skills https://synthedia.substack.com/p/google-funsearch-shows-a-path-for (more thoughts on FunSearch) #AI #math #discoveries
FunSearch: Making new discoveries in mathematical sciences using Large Language Models https://deepmind.google/discover/blog/funsearch-making-new-discoveries-in-mathematical-sciences-using-large-language-models/ #AI #math #discoveries
December 17, 2023 at 12:00 AM
FunSearch: Making new discoveries in mathematical sciences using Large Language Models https://deepmind.google/discover/blog/funsearch-making-new-discoveries-in-mathematical-sciences-using-large-language-models/ #AI #math #discoveries
This release inspired me! I went and read FunSearch, what an awesome and simple idea.
I implemented something inspired by it to optimise agents within my jax "swarm" environment. It works with only 3x1080 Ti running Qwen3-14B, amazing!
Repo github.com/JakeForsey/s...
Best github.com/JakeForsey/s...
I implemented something inspired by it to optimise agents within my jax "swarm" environment. It works with only 3x1080 Ti running Qwen3-14B, amazing!
Repo github.com/JakeForsey/s...
Best github.com/JakeForsey/s...
May 21, 2025 at 8:29 AM
This release inspired me! I went and read FunSearch, what an awesome and simple idea.
I implemented something inspired by it to optimise agents within my jax "swarm" environment. It works with only 3x1080 Ti running Qwen3-14B, amazing!
Repo github.com/JakeForsey/s...
Best github.com/JakeForsey/s...
I implemented something inspired by it to optimise agents within my jax "swarm" environment. It works with only 3x1080 Ti running Qwen3-14B, amazing!
Repo github.com/JakeForsey/s...
Best github.com/JakeForsey/s...
dependence of traditional ones in variational quantum algorithms. Our approach (FunSearch) autonomously discovers hardware-efficient ans\"atze with new features of scalability and system-size-independent number of variational parameters entirely from [2/5 of https://arxiv.org/abs/2505.06347v1]
May 13, 2025 at 6:15 AM
dependence of traditional ones in variational quantum algorithms. Our approach (FunSearch) autonomously discovers hardware-efficient ans\"atze with new features of scalability and system-size-independent number of variational parameters entirely from [2/5 of https://arxiv.org/abs/2505.06347v1]
Even with LLMs you *could* do verifiable output, though I agree that's not how students use ChatGPT. Something like the FunSearch paradigm for example seems promising to me – get an LLM to prove theorems, but in machine-checkable form, not just prose that you have to check yourself for correctness.
FunSearch: Making new discoveries in mathematical sciences using Large Language Models
We introduce FunSearch, a method for searching for “functions” written in computer code, and find new solutions in mathematics and computer science. FunSearch works by pairing a pre-trained LLM,...
deepmind.google
September 30, 2024 at 4:39 PM
Even with LLMs you *could* do verifiable output, though I agree that's not how students use ChatGPT. Something like the FunSearch paradigm for example seems promising to me – get an LLM to prove theorems, but in machine-checkable form, not just prose that you have to check yourself for correctness.
AlphaEvolve: A coding agent for scientific and algorithmic discovery AlphaEvolve, evolves (no pun intended) the seminal method FunSearch introduced in late 2023. Powered by a frontier model rather than a smaller LLM, it leverages evolutionary algorithms...
| Details | Interest | Feed |
| Details | Interest | Feed |
Origin
graphcore-research.github.io
June 3, 2025 at 11:50 PM
Can LLMs be used to discover interpretable models of human and animal behavior?🤔
Turns out: yes!
Thrilled to share our latest preprint where we used FunSearch to automatically discover symbolic cognitive models of behavior.
1/12
Turns out: yes!
Thrilled to share our latest preprint where we used FunSearch to automatically discover symbolic cognitive models of behavior.
1/12
February 10, 2025 at 12:21 PM
Can LLMs be used to discover interpretable models of human and animal behavior?🤔
Turns out: yes!
Thrilled to share our latest preprint where we used FunSearch to automatically discover symbolic cognitive models of behavior.
1/12
Turns out: yes!
Thrilled to share our latest preprint where we used FunSearch to automatically discover symbolic cognitive models of behavior.
1/12
“FunSearch (so called because it searches for mathematical functions, not because it’s fun) “ ... I object to this, searching for mathematical functions can be fun
January 30, 2025 at 11:03 PM
“FunSearch (so called because it searches for mathematical functions, not because it’s fun) “ ... I object to this, searching for mathematical functions can be fun
It's so annoying that people (even within DeepMind) keep saying that OpenAI is ahead "in AI", when DeepMind has models like AlphaFold, AlphaProof, GenCast and GraphCast, FunSearch, etc
AI for Science has such a clearer potential for helping people, but VCs and the media are obsessed by GenAI
AI for Science has such a clearer potential for helping people, but VCs and the media are obsessed by GenAI
September 21, 2024 at 7:15 PM
It's so annoying that people (even within DeepMind) keep saying that OpenAI is ahead "in AI", when DeepMind has models like AlphaFold, AlphaProof, GenCast and GraphCast, FunSearch, etc
AI for Science has such a clearer potential for helping people, but VCs and the media are obsessed by GenAI
AI for Science has such a clearer potential for helping people, but VCs and the media are obsessed by GenAI
#Google #DeepMind FunSearch: Making new discoveries in mathematical sciences using Large Language Models deepmind.google/discover/blo...
FunSearch: Making new discoveries in mathematical sciences using Large Language Models
We introduce FunSearch, a method for searching for “functions” written in computer code, and find new solutions in mathematics and computer science. FunSearch works by pairing a pre-trained LLM,...
deepmind.google
August 1, 2025 at 4:40 PM
#Google #DeepMind FunSearch: Making new discoveries in mathematical sciences using Large Language Models deepmind.google/discover/blo...