Shulou(Shulou.com)12/24 Report--

Intel Research Institute will focus on 31 research findings that will promote future-oriented AI innovation.

Intel Research will showcase a series of valuable and industry-leading AI innovations at NeurIPS 2023. For a wide range of developers, researchers and academics, the world's top conference in the field of AI and computer vision will be held in New Orleans from December 10 to 16.

At NeurIPS 2023, Intel Research Institute will showcase its latest AI research and share Intel's vision of "making AI ubiquitous" with a diverse community of innovators and thought leaders. During the conference, Intel Research Institute will publish 31 papers, including 12 main conference papers and 19 seminar papers, and conduct a technical demonstration at Booth 405. These studies focus on new models, methods and tools developed for the application of AI in science, as well as graph learning, multimodal generation AI, and AI algorithms and optimization techniques for AI use cases such as climate modeling, drug discovery and materials science.

In addition, Intel Research will host the AI accelerated material Discovery (AI4Mat) Seminar on December 15 to provide a platform for AI researchers and materials scientists to jointly discuss how to address the challenges of AI-driven material discovery and development.

Specifically, the research results presented by Intel Research can be divided into the following categories, each of which has some influence in its field:

1.AI-driven Scientific Research (AI for Science)

Brain coding model: a model developed with researchers at the University of Texas at Austin to help predict brain responses and understand the brain's multimodal processing capabilities.

ClimateSet: a large-scale climate model dataset developed jointly with the Quebec Institute of artificial Intelligence (Mila) for machine learning research to quickly predict new climate change scenarios (climate change scenarios) and to lay the foundation for innovative climate applications in machine learning.

HoneyBee: an advanced large language model developed jointly with the Quebec Institute of artificial Intelligence to help researchers understand materials science faster.

two。 Multi-mode generating AI

COCO-Counterfactuals: a multimodal technique for generating synthetic counterfactual (counterfactual) data, which can reduce incorrect statistical deviations in pre-trained multimodal models and help improve the performance of AI models in performing downstream tasks such as image retrieval and image recognition.

LDM3D-VR: a potential Diffusion Model (latent diffusion model) for 3D virtual reality that simplifies 3D video generation in AI applications.

CorresNeRF: an image rendering method that uses nerve radiation field (neural radiance fields) to reconstruct 3D representation of scene from 2D image.

3. Improve AI performanc

Diffpack: a generative AI method for protein modeling, which helps to ensure that the generated 3D structure can reflect the real structural characteristics of the protein.

InstaTune: a method of generating a supernetwork (super-network) during the tuning phase that reduces the total amount of time and computing resources required for network attached storage (NAS).

4. Graph learning

A*Net: a leading path-based knowledge graph reasoning method with millions of datasets, which can extend the dataset beyond the limits of computing and improve the accuracy of large language models.

ULTRA: the industry's leading basic reasoning model for knowledge graphs, as well as new methods for learning general and transferable graph representations and their relationships.

Perfograph: a new program representation based on compiler graph, which can capture numerical information and composite data structures, and improve the ability of machine learning methods to deduce programming languages.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.