Curated Papers > In recent years, there has been significant research in "AI Drug Discovery" with a high impact factor (IF≥30) over the past five years.

Scientific discovery in the age of artificial intelligence.

Literature Information

DOI	10.1038/s41586-023-06221-2
PMID	37532811
Journal	Nature
Impact Factor	48.5
JCR Quartile	Q1
Publication Year	2023
Times Cited	245
Keywords	Artificial Intelligence, Scientific Discovery, Self-Supervised Learning, Geometric Deep Learning, Generative AI
Literature Type	Journal Article, Review, Research Support, N.I.H., Extramural, Research Support, U.S. Gov't, Non-P.H.S., Research Support, Non-U.S. Gov't
ISSN	0028-0836
Pages	47-60
Issue	620(7972)
Authors	Hanchen Wang, Tianfan Fu, Yuanqi Du, Wenhao Gao, Kexin Huang, Ziming Liu, Payal Chandak, Shengchao Liu, Peter Van Katwyk, Andreea Deac, Anima Anandkumar, Karianne Bergen, Carla P Gomes, Shirley Ho, Pushmeet Kohli, Joan Lasenby, Jure Leskovec, Tie-Yan Liu, Arjun Manrai, Debora Marks, Bharath Ramsundar, Le Song, Jimeng Sun, Jian Tang, Petar Veličković, Max Welling, Linfeng Zhang, Connor W Coley, Yoshua Bengio, Marinka Zitnik

TL;DR

This paper explores the integration of artificial intelligence (AI) in scientific discovery, highlighting advancements like self-supervised learning and geometric deep learning that enhance research efficiency and accuracy. It emphasizes the potential of generative AI in creating new designs, such as drugs and proteins, while addressing the ongoing challenges of data quality and the need for improved understanding of AI tools among researchers, thereby underscoring critical areas for future innovation in AI.

Search for more papers on MaltSci.com

Artificial Intelligence · Scientific Discovery · Self-Supervised Learning · Geometric Deep Learning · Generative AI

Abstract

Artificial intelligence (AI) is being increasingly integrated into scientific discovery to augment and accelerate research, helping scientists to generate hypotheses, design experiments, collect and interpret large datasets, and gain insights that might not have been possible using traditional scientific methods alone. Here we examine breakthroughs over the past decade that include self-supervised learning, which allows models to be trained on vast amounts of unlabelled data, and geometric deep learning, which leverages knowledge about the structure of scientific data to enhance model accuracy and efficiency. Generative AI methods can create designs, such as small-molecule drugs and proteins, by analysing diverse data modalities, including images and sequences. We discuss how these methods can help scientists throughout the scientific process and the central issues that remain despite such advances. Both developers and users of AI toolsneed a better understanding of when such approaches need improvement, and challenges posed by poor data quality and stewardship remain. These issues cut across scientific disciplines and require developing foundational algorithmic approaches that can contribute to scientific understanding or acquire it autonomously, making them critical areas of focus for AI innovation.

MaltSci.com AI Research Service

Intelligent ReadingAnswer any question about the paper and explain complex charts and formulas

Locate StatementsFind traces of a specific claim within the paper

Add to KBasePerform data extraction, report drafting, and advanced knowledge mining

Primary Questions Addressed

1. How can self-supervised learning be further utilized to improve the efficiency of hypothesis generation in scientific research?
2. What are the implications of geometric deep learning for specific scientific disciplines, such as biology or materials science?
3. In what ways can generative AI methods enhance the drug discovery process beyond current capabilities?
4. What strategies can researchers employ to mitigate the challenges posed by poor data quality in AI-driven scientific discovery?
5. How can interdisciplinary collaboration improve the development of foundational algorithmic approaches in AI for scientific understanding?

Key Findings

Research Background and Objectives

Artificial intelligence (AI) is increasingly utilized in scientific research to enhance and expedite various processes, including hypothesis generation, experimental design, data collection, and interpretation. The objective of this study is to explore recent advancements in AI, particularly self-supervised learning and geometric deep learning, and their implications for scientific discovery.

Main Methods/Materials/Experimental Design

The study reviews key AI methodologies that have emerged over the past decade, emphasizing the following:

1. Self-Supervised Learning: This approach allows models to be trained on large amounts of unlabelled data, improving their ability to learn patterns without extensive human annotation.

2. Geometric Deep Learning: This technique incorporates knowledge about the structure of scientific data to enhance model performance, particularly in tasks that involve complex relationships among data points.

3. Generative AI Methods: These methods can synthesize new designs, such as small-molecule drugs and proteins, by analyzing diverse data types, including images and sequences.

Technical Workflow

Key Results and Findings

• The integration of self-supervised learning and geometric deep learning has led to significant improvements in model accuracy and efficiency.
• Generative AI techniques have demonstrated potential in creating novel compounds and proteins, thereby streamlining the drug discovery process.
• Despite these advancements, challenges remain, particularly concerning data quality and the need for better stewardship in AI applications.

Main Conclusions/Significance/Innovation

The study highlights the transformative potential of AI in scientific research, emphasizing that:

• AI can significantly augment traditional methods by enabling more rapid and accurate scientific inquiry.
• There is a critical need for researchers to understand the limitations and appropriate contexts for AI applications to avoid pitfalls associated with poor data quality.
• Foundational algorithmic approaches must be developed to improve scientific understanding and facilitate autonomous learning.

Research Limitations and Future Directions

• The study acknowledges limitations related to the generalizability of AI methods across different scientific disciplines.
• Future research should focus on:
  • Improving data quality and management practices to enhance AI model performance.
  • Developing better algorithms that can adapt to various scientific contexts.
  • Encouraging interdisciplinary collaboration to address the multifaceted challenges posed by AI in scientific discovery.

Aspect	Details
AI Techniques Reviewed	Self-Supervised Learning, Geometric Deep Learning, Generative AI
Key Improvements	Enhanced model accuracy, efficient hypothesis generation
Remaining Challenges	Data quality, need for better stewardship
Future Research Focus	Algorithm development, interdisciplinary collaboration

References

1. Deep learning. - Yann LeCun;Yoshua Bengio;Geoffrey Hinton - Nature (2015)
2. A robotic platform for flow synthesis of organic compounds informed by AI planning. - Connor W Coley;Dale A Thomas;Justin A M Lummiss;Jonathan N Jaworski;Christopher P Breen;Victor Schultz;Travis Hart;Joshua S Fishman;Luke Rogers;Hanyu Gao;Robert W Hicklin;Pieter P Plehiers;Joshua Byington;John S Piotti;William H Green;A John Hart;Timothy F Jamison;Klavs F Jensen - Science (New York, N.Y.) (2019)
3. Advancing mathematics by guiding human intuition with AI. - Alex Davies;Petar Veličković;Lars Buesing;Sam Blackwell;Daniel Zheng;Nenad Tomašev;Richard Tanburn;Peter Battaglia;Charles Blundell;András Juhász;Marc Lackenby;Geordie Williamson;Demis Hassabis;Pushmeet Kohli - Nature (2021)
4. Highly accurate protein structure prediction with AlphaFold. - John Jumper;Richard Evans;Alexander Pritzel;Tim Green;Michael Figurnov;Olaf Ronneberger;Kathryn Tunyasuvunakool;Russ Bates;Augustin Žídek;Anna Potapenko;Alex Bridgland;Clemens Meyer;Simon A A Kohl;Andrew J Ballard;Andrew Cowie;Bernardino Romera-Paredes;Stanislav Nikolov;Rishub Jain;Jonas Adler;Trevor Back;Stig Petersen;David Reiman;Ellen Clancy;Michal Zielinski;Martin Steinegger;Michalina Pacholska;Tamas Berghammer;Sebastian Bodenstein;David Silver;Oriol Vinyals;Andrew W Senior;Koray Kavukcuoglu;Pushmeet Kohli;Demis Hassabis - Nature (2021)
5. A Deep Learning Approach to Antibiotic Discovery. - Jonathan M Stokes;Kevin Yang;Kyle Swanson;Wengong Jin;Andres Cubillos-Ruiz;Nina M Donghia;Craig R MacNair;Shawn French;Lindsey A Carfrae;Zohar Bloom-Ackermann;Victoria M Tran;Anush Chiappino-Pepe;Ahmed H Badran;Ian W Andrews;Emma J Chory;George M Church;Eric D Brown;Tommi S Jaakkola;Regina Barzilay;James J Collins - Cell (2020)
6. The art and practice of structure-based drug design: a molecular modeling perspective. - R S Bohacek;C McMartin;W C Guida - Medicinal research reviews (1996)
7. Autonomous navigation of stratospheric balloons using reinforcement learning. - Marc G Bellemare;Salvatore Candido;Pablo Samuel Castro;Jun Gong;Marlos C Machado;Subhodeep Moitra;Sameera S Ponda;Ziyu Wang - Nature (2020)
8. Unsupervised word embeddings capture latent knowledge from materials science literature. - Vahe Tshitoyan;John Dagdelen;Leigh Weston;Alexander Dunn;Ziqin Rong;Olga Kononova;Kristin A Persson;Gerbrand Ceder;Anubhav Jain - Nature (2019)
9. Deep Potential Molecular Dynamics: A Scalable Model with the Accuracy of Quantum Mechanics. - Linfeng Zhang;Jiequn Han;Han Wang;Roberto Car;Weinan E - Physical review letters (2018)
10. Reducing the dimensionality of data with neural networks. - G E Hinton;R R Salakhutdinov - Science (New York, N.Y.) (2006)

Literatures Citing This Work

1. AI tools as science policy advisers? The potential and the pitfalls. - Chris Tyler;K L Akerlof;Alessandro Allegra;Zachary Arnold;Henriette Canino;Marius A Doornenbal;Josh A Goldstein;David Budtz Pedersen;William J Sutherland - Nature (2023)
2. PLOS Biology at 20: Reflecting on the road we've traveled. - Hemai Parthasarathy;Theodora Bloom;Emma Ganley - PLoS biology (2023)
3. A focus on harnessing big data and artificial intelligence: revolutionizing drug discovery from traditional Chinese medicine sources. - Mingyu Li;Jian Zhang - Chemical science (2023)
4. From beasts to bytes: Revolutionizing zoological research with artificial intelligence. - Yu-Juan Zhang;Zeyu Luo;Yawen Sun;Junhao Liu;Zongqing Chen - Zoological research (2023)
5. Advances in Designing Essential Oil Nanoformulations: An Integrative Approach to Mathematical Modeling with Potential Application in Food Preservation. - Monisha Soni;Arati Yadav;Akash Maurya;Somenath Das;Nawal Kishore Dubey;Abhishek Kumar Dwivedy - Foods (Basel, Switzerland) (2023)
6. Hypotheses devised by AI could find 'blind spots' in research. - Matthew Hutson - Nature (2023)
7. Artificial intelligence for science-bridging data to wisdom. - Yongjun Xu;Fei Wang;Zhulin An;Qi Wang;Zhao Zhang - Innovation (Cambridge (Mass.)) (2023)
8. ChatGPT Research Group for Optimizing the Crystallinity of MOFs and COFs. - Zhiling Zheng;Oufan Zhang;Ha L Nguyen;Nakul Rampal;Ali H Alawadhi;Zichao Rong;Teresa Head-Gordon;Christian Borgs;Jennifer T Chayes;Omar M Yaghi - ACS central science (2023)
9. Knowledge-driven learning, optimization, and experimental design under uncertainty for materials discovery. - Xiaoning Qian;Byung-Jun Yoon;Raymundo Arróyave;Xiaofeng Qian;Edward R Dougherty - Patterns (New York, N.Y.) (2023)
10. Rational Design of Flexible Mechanical Force Sensors for Healthcare and Diagnosis. - Hang Zhang;Yihui Zhang - Materials (Basel, Switzerland) (2023)

... (235 more literatures)

---

© 2025 MaltSci - We reshape scientific research with AI technology