Uday Singh Saini
Research Subjects
Research Interests
Uday Singh Saini has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
- Chin-Chia Michael Yeh
Visa Research, Foster City, CA, USA
3 shared publications - Junpeng Wang
Visa Research, Foster City, CA, USA
3 shared publications - Xin Dai
Visa Research, Foster City, CA, USA
3 shared publications - Evangelos E. Papalexakis
University of California, Riverside, Riverside, CA, USA
2 shared publications - Huiyuan Chen
Case Western Reserve University
2 shared publications - Jiarui Sun
Visa Research, Foster City, CA, USA
2 shared publications
Latest Publications
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 7185-7196
TransactionGPT: Toward Foundational Transaction Modeling
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 7185-7196
Open AccessTransactionGPT: Toward Foundational Transaction Modeling
- Open Access
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2518-2527
TREASURE: A Transformer-Based Foundation Model for High-Volume Transaction Understanding
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2518-2527
Open AccessTREASURE: A Transformer-Based Foundation Model for High-Volume Transaction Understanding
- Open Access
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 14-23
Neural Additive Tensor Decomposition for Sparse Tensors
Given a sparse tensor, how can we accurately capture complex latent structures inherent in the tensor while maintaining the interpretability of those structures? Tensor decomposition is a fundamental technique for analyzing tensors. Classical tensor models provide multi-linear structures that are easy to interpret, but have limitations in capturing complex structures present in real-world sparse tensors. Recent neural tensor models have extended the capabilities of classical tensor models in capturing complex structures within the data. However, this has come at the cost of interpretability: neural tensor models entangle interactions across and within latent structures in a black-box manner, making it difficult to readily understand the discovered structures. Understanding these structures, however, is crucial in applications such as healthcare, which requires transparency in critical decision-making processes. To overcome this major limitation and bridge the gap between the classical multi-linear models and neural tensor models, we propose Neural Additive Tensor Decomposition (NeAT), an accurate and interpretable neural tensor model for sparse tensors. The main idea of NeAT is to apply neural networks to each latent component in an additive fashion. This not only captures diverse patterns and complex structures in sparse tensors, but also provides a direct and intuitive interpretation of the structures by being close to the multi-linear tensor model. We conduct extensive experiments on six large real-world sparse tensors. NeAT outperforms the state-of-the-art neural tensor models in link prediction, surpassing a linear tensor model by 10% and the second-best neural tensor model by 4%, in accuracy. Through ablation studies, we explore various model designs for NeAT to identify key factors that impact generalization. Finally, we evaluate qualitatively and quantitatively latent patterns discovered by NeAT, demonstrating how to analyze the discovered latent patterns in real data obtained from NeAT.
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 14-23
Open AccessNeural Additive Tensor Decomposition for Sparse Tensors
Given a sparse tensor, how can we accurately capture complex latent structures inherent in the tensor while maintaining the interpretability of those structures? Tensor decomposition is a fundamental technique for analyzing tensors. Classical tensor models provide multi-linear structures that are easy to interpret, but have limitations in capturing complex structures present in real-world sparse tensors. Recent neural tensor models have extended the capabilities of classical tensor models in capturing complex structures within the data. However, this has come at the cost of interpretability: neural tensor models entangle interactions across and within latent structures in a black-box manner, making it difficult to readily understand the discovered structures. Understanding these structures, however, is crucial in applications such as healthcare, which requires transparency in critical decision-making processes. To overcome this major limitation and bridge the gap between the classical multi-linear models and neural tensor models, we propose Neural Additive Tensor Decomposition (NeAT), an accurate and interpretable neural tensor model for sparse tensors. The main idea of NeAT is to apply neural networks to each latent component in an additive fashion. This not only captures diverse patterns and complex structures in sparse tensors, but also provides a direct and intuitive interpretation of the structures by being close to the multi-linear tensor model. We conduct extensive experiments on six large real-world sparse tensors. NeAT outperforms the state-of-the-art neural tensor models in link prediction, surpassing a linear tensor model by 10% and the second-best neural tensor model by 4%, in accuracy. Through ablation studies, we explore various model designs for NeAT to identify key factors that impact generalization. Finally, we evaluate qualitatively and quantitatively latent patterns discovered by NeAT, demonstrating how to analyze the discovered latent patterns in real data obtained from NeAT.
- Open Access
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 3919-3930
RPMixer: Shaking Up Time Series Forecasting with Random Projections for Large Spatial-Temporal Data
Spatial-temporal forecasting systems play a crucial role in addressing numerous real-world challenges. In this paper, we investigate the potential of addressing spatial-temporal forecasting problems using general time series forecasting models, i.e., models that do not leverage the spatial relationships among the nodes. We propose a all-Multi-Layer Perceptron (all-MLP) time series forecasting architecture called RPMixer. The all-MLP architecture was chosen due to its recent success in time series forecasting benchmarks. Furthermore, our method capitalizes on the ensemble-like behavior of deep neural networks, where each individual block within the network behaves like a base learner in an ensemble model, particularly when identity mapping residual connections are incorporated. By integrating random projection layers into our model, we increase the diversity among the blocks' outputs, thereby improving the overall performance of the network. Extensive experiments conducted on the largest spatial-temporal forecasting benchmark datasets demonstrate that the proposed method outperforms 14 alternative methods.
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 3919-3930
Open AccessRPMixer: Shaking Up Time Series Forecasting with Random Projections for Large Spatial-Temporal Data
Spatial-temporal forecasting systems play a crucial role in addressing numerous real-world challenges. In this paper, we investigate the potential of addressing spatial-temporal forecasting problems using general time series forecasting models, i.e., models that do not leverage the spatial relationships among the nodes. We propose a all-Multi-Layer Perceptron (all-MLP) time series forecasting architecture called RPMixer. The all-MLP architecture was chosen due to its recent success in time series forecasting benchmarks. Furthermore, our method capitalizes on the ensemble-like behavior of deep neural networks, where each individual block within the network behaves like a base learner in an ensemble model, particularly when identity mapping residual connections are incorporated. By integrating random projection layers into our model, we increase the diversity among the blocks' outputs, thereby improving the overall performance of the network. Extensive experiments conducted on the largest spatial-temporal forecasting benchmark datasets demonstrate that the proposed method outperforms 14 alternative methods.
- Open Access
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 2036-2048
CARL-G: Clustering-Accelerated Representation Learning on Graphs
Self-supervised learning on graphs has made large strides in achieving great performance in various downstream tasks. However, many state-of-the-art methods suffer from a number of impediments, which prevent them from realizing their full potential. For instance, contrastive methods typically require negative sampling, which is often computationally costly. While non-contrastive methods avoid this expensive step, most existing methods either rely on overly complex architectures or dataset-specific augmentations. In this paper, we ask: Can we borrow from classical unsupervised machine learning literature in order to overcome those obstacles? Guided by our key insight that the goal of distance-based clustering closely resembles that of contrastive learning: both attempt to pull representations of similar items together and dissimilar items apart. As a result, we propose CARL-G - a novel clustering-based framework for graph representation learning that uses a loss inspired by Cluster Validation Indices (CVIs), i.e., internal measures of cluster quality (no ground truth required). CARL-G is adaptable to different clustering methods and CVIs, and we show that with the right choice of clustering method and CVI, CARL-G outperforms node classification baselines on 4/5 datasets with up to a 79× training speedup compared to the best-performing baseline. CARL-G also performs at par or better than baselines in node clustering and similarity search tasks, training up to 1,500× faster than the best-performing baseline. Finally, we also provide theoretical foundations for the use of CVI-inspired losses in graph representation learning.
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 2036-2048
Open AccessCARL-G: Clustering-Accelerated Representation Learning on Graphs
Self-supervised learning on graphs has made large strides in achieving great performance in various downstream tasks. However, many state-of-the-art methods suffer from a number of impediments, which prevent them from realizing their full potential. For instance, contrastive methods typically require negative sampling, which is often computationally costly. While non-contrastive methods avoid this expensive step, most existing methods either rely on overly complex architectures or dataset-specific augmentations. In this paper, we ask: Can we borrow from classical unsupervised machine learning literature in order to overcome those obstacles? Guided by our key insight that the goal of distance-based clustering closely resembles that of contrastive learning: both attempt to pull representations of similar items together and dissimilar items apart. As a result, we propose CARL-G - a novel clustering-based framework for graph representation learning that uses a loss inspired by Cluster Validation Indices (CVIs), i.e., internal measures of cluster quality (no ground truth required). CARL-G is adaptable to different clustering methods and CVIs, and we show that with the right choice of clustering method and CVI, CARL-G outperforms node classification baselines on 4/5 datasets with up to a 79× training speedup compared to the best-performing baseline. CARL-G also performs at par or better than baselines in node clustering and similarity search tasks, training up to 1,500× faster than the best-performing baseline. Finally, we also provide theoretical foundations for the use of CVI-inspired losses in graph representation learning.