Article Introduction：Editor | Ziluo's traditional material research and development model mainly relies on "trial and error" experimental methods or accidental discoveries, and its research and development process usually takes 10-20 years. Data-driven methods based on machine learning (ML) can accelerate the design of new materials for clean energy technologies. However, its practical application in materials research is still limited due to the lack of large-scale high-fidelity experimental databases. Recently, research teams from the Pacific Northwest National Laboratory and Argonne National Laboratory in the United States designed a highly automated workflow that combines a high-throughput experimental platform with the most advanced active learning algorithms to effectively screen for anolyte electrolytes. Binary organic solvent for optimal solubility. The goal of this research is to improve the performance and stability of energy storage systems to promote renewable energy

2024-04-10 comment 0  523

Article Introduction：According to news from this site on September 3, accurate prediction of lithium battery life is crucial for the normal operation of electrical equipment. However, accurate prediction of battery life faces challenges due to the nonlinearity of the battery capacity degradation process and the uncertainty of operating conditions. The Chinese Academy of Sciences stated that the team of researcher Chen Zhongwei and associate researcher Mao Zhiyu from the Power Battery and System Research Department of the National Key Laboratory of Energy Catalytic Conversion of the Dalian Institute of Chemical Physics, together with Professor Feng Jiangtao of Xi'an Jiaotong University, have made progress in battery health management research. Relevant research results have been published in the Journal of Transportation Electrochemistry of the Institute of Electrical and Electronics Engineers (DOI: 10.1109/TTE.2024.3434553 attached to this site). 1. According to reports, the research team developed a new deep learning model

2024-09-03 comment 0  430

Article Introduction：ACMCCS2023, the international authoritative academic conference in the computer field, opened on November 26 in Copenhagen, Denmark. The paper "Hopper: Interpretative Fuzzing for Libraries" by the Tencent Security Big Data Laboratory team was included in the conference. Yesterday, laboratory researcher Xie Yuxuan was invited to attend the conference to share the theme. This research proposes an interpretive fuzz testing method and demonstrates how to use dynamic feedback to learn constraints inside and outside the API to achieve automated code generation. Through this method, without any external expert knowledge, it is possible to generate valid and usable code calling methods and exploit these codes to exploit vulnerabilities. The goal of this research method is to solve the human need for fuzz testing

2023-11-29 comment 0  1476

Article Introduction：[CNMO News] On November 14, according to Xinhua News Agency, the team of professors Luo Yi, Jiang Jun, and Shang Weiwei of the University of Science and Technology of China recently cooperated with researcher Zhang Zhe of the Deep Space Exploration Laboratory and others to use the intelligent robot "Machine Chemist" , successfully developed a new catalyst using Martian meteorites, providing a high-efficiency, low-energy solution for using water on Mars to produce oxygen, and exploring a new way to develop chemicals using local materials in galaxies outside the Earth. Today, the internationally renowned academic journal Nature Synthesis published this research result. According to reports, researchers from the University of Science and Technology of China and the Deep Space Exploration Laboratory collaborated to use their self-developed intelligent robot "Machine Chemist" to analyze and extract components from Martian meteorites, successfully

2023-11-14 comment 0  856

Article Introduction：Editor | ScienceAI Recently, the research group of Professor Hong Liang from the Institute of Natural Sciences/School of Physics and Astronomy/Zhangjiang Advanced Research Institute/School of Pharmacy of Shanghai Jiao Tong University, and Tan Pan, a young researcher from the Shanghai Artificial Intelligence Laboratory, achieved important breakthroughs in protein mutation-property prediction. . This work adopts a new training strategy, which greatly improves the performance of traditional protein pre-trained large models in mutation-property prediction using very little wet experimental data. The research results are titled "Enhancing the efficiency of protein language models with minimal wet-labdata throughfew-shot learning" and were published in

2024-07-11 comment 0  913

Article Introduction：The AIxiv column is a column where this site publishes academic and technical content. In the past few years, the AIxiv column of this site has received more than 2,000 reports, covering top laboratories from major universities and companies around the world, effectively promoting academic exchanges and dissemination. If you have excellent work that you want to share, please feel free to contribute or contact us for reporting. Submission email: liyazhou@jiqizhixin.com; zhaoyunfeng@jiqizhixin.com This article was completed by Shanghai Artificial Intelligence Laboratory, Dalian University of Technology and University of Science and Technology of China. Corresponding author: Shao Jing, graduated from the Multimedia Laboratory MMLab of the Chinese University of Hong Kong with a Ph.D., and is currently the head of the large model security team of Pujiang National Laboratory, leading the research on large models.

2024-06-14 comment 0  454

Article Introduction：Editor | Dead Leaf Butterfly Large-scale language models have greatly enhanced scientists' ability to understand biology and chemistry, but reliable methods for structure-based drug discovery, quantum chemistry, and structural biology remain few. Accurate biomolecule-ligand interaction datasets are urgently needed for large language models. In order to solve this problem, researchers from the Institute of Biology of the Helmholtz Research Center München and the Technical University of Munich proposed MISATO. This is a data set that combines quantum mechanical (QM) properties of small molecules with associated molecular dynamics (MD) simulations of approximately 20,000 experimental protein-ligand complexes, and extensive validation of experimental data. Starting from the existing experimental structure, the researchers used semi-empirical quantum mechanics to systematically improve these

2024-06-01 comment 0  462

Article Introduction：Chemical reactions are the basis of drug design and organic chemistry research. There is a growing need among the research community for a large-scale deep learning framework that can effectively capture the fundamental rules of chemical reactions. Recently, a research team from Peking University and Wangshi Intelligence proposed a new method to bridge the gap between reaction-based molecular pre-training and generation tasks. Inspired by the mechanisms of organic chemistry, researchers developed a new pre-training framework that enables it to incorporate inductive bias into models. This proposed framework achieves state-of-the-art results when performing challenging downstream tasks. By leveraging knowledge of chemistry, the framework overcomes the limitations of current molecular generation models that rely on a small number of reaction templates. In extensive experiments, the model generated high-quality synthesizable drug-like structures. Overall, the model

2023-12-14 comment 0  627

Article Introduction：The AIxiv column is a column where this site publishes academic and technical content. In the past few years, the AIxiv column of this site has received more than 2,000 reports, covering top laboratories from major universities and companies around the world, effectively promoting academic exchanges and dissemination. If you have excellent work that you want to share, please feel free to contribute or contact us for reporting. Submission email: liyazhou@jiqizhixin.com; zhaoyunfeng@jiqizhixin.com The main author of this article is Huang Yichong. Huang Yichong is a doctoral student at the Social Computing and Information Retrieval Research Center of Harbin Institute of Technology and an intern at Pengcheng Laboratory, studying under Professor Qin Bing and Professor Feng Xiaocheng. Research directions include ensemble learning of large language models, multi-language large models, correlation theory

2024-07-19 comment 0  1222

Article Introduction：The AIxiv column is a column where this site publishes academic and technical content. In the past few years, the AIxiv column of this site has received more than 2,000 reports, covering top laboratories from major universities and companies around the world, effectively promoting academic exchanges and dissemination. If you have excellent work that you want to share, please feel free to contribute or contact us for reporting. Submission email: liyazhou@jiqizhixin.com; zhaoyunfeng@jiqizhixin.com The research team of Shanghai Jiao Tong University’s Generative Artificial Intelligence Laboratory (GAIRLab)’s main research directions are: large model training, alignment and evaluation. Team homepage: https://plms.ai/AI technology is changing with each passing day. Recently, Anthr

2024-06-24 comment 0  1113

Article Introduction：Apple recently announced that it will expand the scale of its applied research laboratory in China to better support product manufacturing and research and development. It is reported that Apple plans to strengthen the Shanghai research center and provide more comprehensive support to ensure that all product lines reach higher levels in terms of reliability, quality and material analysis. Apple plans to set up a new applied research laboratory in Shenzhen. This move is seen as a strategic layout to introduce VisionPro products into the Chinese market. The company said it will strengthen employee support in the region and deepen cooperation with local suppliers. This new laboratory will focus on improving the testing and research capabilities of products such as iPhone, iPad and Apple Vision Pro. According to the editor's investigation, Apple has already

2024-03-14 comment 0  471

Article Introduction：Recently, a team of professors Luo Yi, Jiang Jun and Shang Weiwei from the University of Science and Technology of China collaborated with researcher Zhang Zhe from the Deep Space Exploration Laboratory to successfully develop a new type of catalyst for use on Mars using an intelligent robot "machine chemist". Use water to produce oxygen. This research provides an efficient, low-energy solution and opens up a new path for the development of chemicals from local materials in extraterrestrial galaxies. The research results were published in the internationally renowned academic journal "Nature·Synthesis" on November 14. Editor: Li Hengyi AI synthesis anchor Technical support: iFlytek

2023-11-14 comment 0  688

Article Introduction：The AIxiv column is a column where this site publishes academic and technical content. In the past few years, the AIxiv column of this site has received more than 2,000 reports, covering top laboratories from major universities and companies around the world, effectively promoting academic exchanges and dissemination. If you have excellent work that you want to share, please feel free to contribute or contact us for reporting. Submission email: liyazhou@jiqizhixin.com; zhaoyunfeng@jiqizhixin.com The research team of Shanghai Jiao Tong University’s Generative Artificial Intelligence Laboratory (GAIRLab)’s main research directions are: large model training, alignment and evaluation. Team homepage: https://plms.ai/ In the next 20 years, AI is expected to surpass humans

2024-06-20 comment 0  653

Article Introduction：This article is a study on improving the scalability of zero-order optimization. The code has been open source and the paper has been accepted by ICLR2024. Today I’d like to introduce a paper titled “DeepZero: ScalingupZeroth-OrderOptimization for DeepModelTraining”, a collaboration between Michigan State University and Lawrence Livermore National Laboratory. This paper was recently accepted by the ICLR2024 conference, and the research team has made the code open source. The main goal of this paper is to extend zero-order optimization techniques in deep learning model training. Zero-order optimization is an optimization method that does not rely on gradient information. It can better handle high-dimensional parameter spaces and complex models.

2024-02-15 comment 0  720

Article Introduction：The AIxiv column is a column where this site publishes academic and technical content. In the past few years, the AIxiv column of this site has received more than 2,000 reports, covering top laboratories from major universities and companies around the world, effectively promoting academic exchanges and dissemination. If you have excellent work that you want to share, please feel free to contribute or contact us for reporting. Submission email: liyazhou@jiqizhixin.com; zhaoyunfeng@jiqizhixin.com This article was completed by HMILab. Relying on the two major platforms of Peking University’s National Engineering Research Center for Video and Visual Technology and the National Key Laboratory of Multimedia Information Processing, HMILab has long been engaged in research in the direction of machine learning, multi-modal learning and embodied intelligence. This work No.

2024-07-16 comment 0  488

Article Introduction：The TinyLLaVA+ project was jointly created by Professor Wu Ji’s team from the Multimedia Signal and Intelligent Information Processing Laboratory (MSIIP) of the Department of Electronics, Tsinghua University, and Professor Huang Lei’s team from the School of Artificial Intelligence, Beihang University. The MSIIP Laboratory of Tsinghua University has long been committed to research fields such as intelligent medical care, natural language processing and knowledge discovery, and multi-modality. The Beijing Airlines team has long been committed to research fields such as deep learning, multi-modality, and computer vision. The goal of the TinyLLaVA+ project is to develop a small cross-language intelligent assistant with multi-modal capabilities such as language understanding, question and answer, and dialogue. The project team will give full play to their respective advantages, jointly overcome technical problems, and realize the design and development of intelligent assistants. This will provide opportunities for intelligent medical care, natural language processing and knowledge

2024-06-08 comment 0  498

Article Introduction：Large models can form your own understanding of the real world! An MIT study found that as a model becomes more capable, its understanding of reality may go beyond simple imitation. For example, if the large model has never smelled a smell, does that mean it cannot understand smells? Research has found that it can spontaneously simulate some concepts for easy understanding. This research means that large models are expected to have a deeper understanding of language and the world in the future, and the paper has been accepted by ICML24, the top conference. The authors of this paper are Chinese doctoral student Charles Jin and his supervisor Professor Martin Rinard from MIT Computer and Artificial Intelligence Laboratory (CSAIL). In the study, the author asked the large model to only learn the code text, and found that the model gradually grasped the meaning behind it. Rin

2024-08-17 comment 0  707

Article Introduction：Editor | Structural prediction of radish skin protein complexes plays an important role in drug development, antibody design and other applications. However, due to limited prediction accuracy, prediction results are often inconsistent with experimental results. A research team from Peking University, Changping Laboratory, and Harvard University proposes ColabDock, a general framework that employs deep learning structural prediction models to integrate different forms and sources of experimental constraints without further large-scale retraining or fine-tuning. . ColabDock outperforms HADDOCK and ClusPro, which use AlphaFold2 as a structure prediction model, not only in complex structure predictions with simulated residues and surface constraints, but also in NMR chemical shift perturbations and covalent labeling.

2024-08-08 comment 0  464

Article Introduction：Does a large language model that can only "read books" have the visual perception of the real world? By modeling the relationships between strings, what exactly can a language model learn about the visual world? Recently, researchers at the MIT Computer Science and Artificial Intelligence Laboratory (MITCSAIL) evaluated language models, focusing on their visual capabilities. They tested the model's ability by asking it to generate and recognize increasingly complex visual concepts, from simple shapes and objects to complex scenes. The researchers also showed how to train a preliminary visual representation learning system using a text-only model. With this research, they laid the foundation for further development and improvement of visual representation learning systems. Paper link: https://arxiv.

2024-02-01 comment 0  1088

Article Introduction：At present, although great research progress has been made in human body pose and shape estimation (EHPS, Expressive Human Pose and Shapeestimation), the most advanced methods are still limited by the limitations of the training data set. Recently, Nanyang Technological University S-Lab, SenseTime Researchers from Science and Technology, Shanghai Artificial Intelligence Laboratory, University of Tokyo and IDEA Research Institute proposed for the first time the large-scale motion capture model SMPLer-X for human body posture and body size estimation tasks. The study used up to 4.5 million instances from different data sources to train the model, achieving top performance on 7 key lists. SMPLer-X not only captures body movements, but also outputs

2023-10-30 comment 0  666