新型FoundationModels及應(yīng)用
訓(xùn)練和推理加速
可解釋性等理論研究
Virtual Agent
Physical Agent
2050 RESEARCH
為昆侖在通用人工智能(AGI)的中長期發(fā)展中提供基礎(chǔ)研究維度上的支持,增強自主創(chuàng)新,加速實現(xiàn)AGI的步伐,同時專注于AGI相關(guān)課題的科技創(chuàng)新,提升企業(yè)的理論與技術(shù)競爭力。
跨越“奇點”、探索未知世界、創(chuàng)造美好未來,成為一家面向未來
世界的卓越科學(xué)研究機構(gòu)!
新加坡工程院院士、AAAI、ACM、IEEE 及 IAPR Fellow
新型FoundationModels及應(yīng)用
訓(xùn)練和推理加速
可解釋性等理論研究
Virtual Agent
Physical Agent
Reinforcement Learning from Diverse Human Preferences
Wanqi Xue, Bo An, Shuicheng Yan, Zhongwen Xu
IJCAI 2024 Conference
August 2024
Keywords: Reinforcement Learning, Human Preferences, Human Feedback, Rewards
Exploring Diffusion Time-steps for Unsupervised Representation Learning
Zhongqi Yue, Jiankun Wang, Qianru Sun, Lei Ji, Eric I-Chao Chang, Hanwang Zhang
ICLR 2024 Conference
May 2024
Keywords: unsupervised representation learning, diffusion model, representation disentanglement, counterfactual generation
Synapse: Trajectory-as-Exemplar Prompting with Memory for Computer Control
Longtao Zheng, Rundong Wang, Xinrun Wang, Bo An
ICLR 2024 Conference
May 2024
Keywords: AI Agents, Large Language Models, Prompting
True Knowledge Comes from Practice: Aligning Large Language Models with Embodied Environments via Reinforcement Learning
Weihao Tan, Wentao Zhang, Shanqi Liu, Longtao Zheng, Xinrun Wang, Bo An
ICLR 2024 Conference
May 2024
Keywords: Reinforcement Learning, Large Language Models, Parameter-Efficient Fine-Tuning
Enhancing Video-Language Representations with Structural Spatio-Temporal Alignment
Hao Fei; Shengqiong Wu; Meishan Zhang; Min Zhang; Tat-Seng Chua; Shuicheng Yan
IEEE Transactions on Pattern Analysis and Machine Intelligence
April 2024
Keywords: Videos, Semantics, Transformers
Skywork
天工系列模型在3.2TB高質(zhì)量多語言和代碼數(shù)據(jù)上進行了預(yù)訓(xùn)練,并開源了模型參數(shù),訓(xùn)練數(shù)據(jù),評估數(shù)據(jù),評估方法。
Vitron
Vitron是一個通用的像素級視覺大語言模型,設(shè)計用于全面理解(感知和推理)、生成、分割(定位和跟蹤)、編輯(修補)靜態(tài)圖像和動態(tài)視頻內(nèi)容。
AgentStudio
AgentStudio是一個開放工具包,覆蓋了構(gòu)建能夠與數(shù)碼世界中的一切進行交互的虛擬代理的整個生命周期。
PointCloudMamba
Point Cloud Mamba 超越了SOTA點云方法PointNeXt,并在ScanObjectNN、ModelNet40和ShapeNetPart數(shù)據(jù)集上達(dá)到了新的SOTA性能。
天工
FOR THE BEST AND THE BRIGHTEST