RI:Small:Exploring Efficient Bayesian Model-Augmentation Techniques for Decomposible Contrastive Representation Learning

RI:Small：探索可分解对比表示学习的高效贝叶斯模型增强技术

• 批准号: 2223292
• 负责人: Changyou Chen
• 金额: $ 41.57万
• 依托单位: SUNY at Buffalo
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-09-01 至 2025-08-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2223292&HistoricalAwards=false
• 关键词: RI; Small; Exploring; Efficient; Bayesian

Modern deep learning models that trains large on web-scale data with a heavy computational load are achieving state-of-the-art performance on many real-world problems. The ability to learning good representations from the large-scale unlabeled data is the key to the success. Although many methods have been developed, their underlying properties and limitations are still not well understood. Contrastive Representation Learning (CRL) is a method of learning to represent the data such that similar data are close to each other while dissimilar data are far apart. This project investigates the limitations of CRL in order to make it more scalable to big data and be appropriately applied to real-world problems such as computer vision. The project will also support the continued development of machine-learning related courses for undergraduate and graduate students at University at Buffalo as well as various outreach activities to expose K-12 students to the field of computer science.This research develops a scalable conditional decomposable contrastive learning framework from the Bayesian principle and extend it to the federated learning and multi-model learning. First, an augmentation technique will be applied to decouple the entanglement of positive-negative samples, leading to a conditional decomposable loss that can be optimized with unbiased stochastic gradients. Second, the technique will be further refined to develop a communication-efficient distributed training framework for CRL, where clients no longer need to explicitly communicate with other clients to fetch other negative samples. Third, the improved CRL technique will be further applied to the emerging field of foundation models for vision-and-language modeling. The project will also result in the dissemination of shared data and benchmarks to the broader AI community, for example through Github, presentation and workshop organization.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

现代深度学习模型在网络规模数据上进行大量计算负载的训练，在许多现实问题上取得了最先进的性能。从大规模未标记数据中学习良好表示的能力是成功的关键。尽管已经开发了许多方法，但它们的基本属性和局限性仍然没有得到很好的理解。对比表示学习（CRL）是一种学习表示数据的方法，使得相似的数据彼此接近，而不同的数据相距很远。该项目研究了 CRL 的局限性，以便使其能够更好地适应大数据，并适当地应用于计算机视觉等现实世界的问题。该项目还将支持布法罗大学本科生和研究生机器学习相关课程的持续开发，以及各种外展活动，让 K-12 学生接触计算机科学领域。这项研究开发了一种可扩展的条件可分解对比学习框架从贝叶斯原理延伸到联邦学习和多模型学习。首先，将应用增强技术来解耦正负样本的纠缠，从而产生可以通过无偏随机梯度进行优化的条件可分解损失。其次，该技术将进一步完善，为 CRL 开发一个通信高效的分布式训练框架，其中客户端不再需要与其他客户端显式通信来获取其他负样本。第三，改进的CRL技术将进一步应用于视觉和语言建模基础模型的新兴领域。该项目还将导致向更广泛的人工智能社区传播共享数据和基准，例如通过 Github、演示和研讨会组织。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力优点和技术进行评估，被认为值得支持。更广泛的影响审查标准。

项目成果

STT: Soft Template Tuning for Few-Shot Adaptation

• DOI: 10.1109/icdmw58026.2022.00122
• 发表时间: 2022-07
• 期刊: 2022 IEEE International Conference on Data Mining Workshops (ICDMW)
• 作者: Ping Yu;Wei Wang-;Chunyuan Li;Ruiyi Zhang;Zhanpeng Jin;Changyou Chen

What Do Audio Transformers Hear? Probing Their Representations For Language Delivery & Structure

• DOI: 10.1109/icdmw58026.2022.00120
• 发表时间: 2022-11
• 期刊: 2022 IEEE International Conference on Data Mining Workshops (ICDMW)
• 作者: Yaman Kumar Singla;Jui Shah;Changyou Chen;R. Shah

Fed-CBS: A Heterogeneity-Aware Client Sampling Mechanism for Federated Learning via Class-Imbalance Reduction

• DOI: 10.48550/arxiv.2209.15245
• 发表时间: 2022-09
• 作者: Jianyi Zhang;Ang Li;Minxue Tang;Jingwei Sun;Xiang Chen;Fan Zhang;Chang Chen;Yiran Chen;H. Li

Persuasion Strategies in Advertisements

广告中的说服策略

• 发表时间: 2023
• 期刊: Proceedings of the AAAI Conference on Artificial Intelligence
• 作者: Yaman Kumar;Rajat Jha;Arunim Gupta;Milan Aggarwal;Aditya Garg;Tushar Malyan;Ayush Bhardwaj;Rajiv Ratn Shah;Balaji Krishnamurthy;Changyou Chen
• 通讯作者: Changyou Chen