Yang Jiao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Yang Jiao has a total of 32 co-authors. Here are the top co-authors by number of shared publications:
- Yan Gao
Amazon, Seattle, USA
5 shared publications Mei Yang
Department of Electrical and Computer Engineering, University of Nevada
3 shared publications- Yi Sun
Amazon, Seattle, USA
3 shared publications - Fan Yang
Wake Forest University
2 shared publications - Mo Weng
School of Life Sciences, University of Nevada, Las Vegas, Las Vegas, USA
2 shared publications - Yetian Chen
Amazon, Seattle, WA, USA
2 shared publications
Latest Publications
- Open Access
- 19 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1812-1822
Multi-modal Relational Item Representation Learning for Inferring Substitutable and Complementary Items
We study the problem of inferring substitutable and complementary items, which underpins applications such as alternative and follow-up purchase suggestions. Existing approaches typically learn from behavior-derived item-item associations using GNNs or leverage item content alone. However, these methods often overlook two key challenges: (i) user behaviors (e.g., co-view/co-purchase) only provide noisy weak supervision, and (ii) behavior signals are long-tailed, leaving many items with sparse associations. We propose MMSC, a self-supervised multi-modal relational representation learning framework that combines a multi-modal foundation model adapted to encode item metadata and a self-supervised denoising module that learns relationship-aware representations from noisy user behaviors, unified by a hierarchical aggregation mechanism. We further use LLM-assisted supervision to mitigate noise in behavior-derived supervision during training. Experiments on five real-world datasets show that MMSC consistently outperforms existing baselines by 26.1% for substitutable and 39.2% for complementary item inference, while remaining effective for cold-start items.
- 19 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1812-1822
Open AccessMulti-modal Relational Item Representation Learning for Inferring Substitutable and Complementary Items
We study the problem of inferring substitutable and complementary items, which underpins applications such as alternative and follow-up purchase suggestions. Existing approaches typically learn from behavior-derived item-item associations using GNNs or leverage item content alone. However, these methods often overlook two key challenges: (i) user behaviors (e.g., co-view/co-purchase) only provide noisy weak supervision, and (ii) behavior signals are long-tailed, leaving many items with sparse associations. We propose MMSC, a self-supervised multi-modal relational representation learning framework that combines a multi-modal foundation model adapted to encode item metadata and a self-supervised denoising module that learns relationship-aware representations from noisy user behaviors, unified by a hierarchical aggregation mechanism. We further use LLM-assisted supervision to mitigate noise in behavior-derived supervision during training. Experiments on five real-world datasets show that MMSC consistently outperforms existing baselines by 26.1% for substitutable and 39.2% for complementary item inference, while remaining effective for cold-start items.
- Open Access
- 20 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 800-811
Learning Attribute as Explicit Relation for Sequential Recommendation
- 20 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 800-811
Open AccessLearning Attribute as Explicit Relation for Sequential Recommendation
- Open Access
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 641-645
Rethinking Sequential Relationships: Improving Sequential Recommenders with Inter-Sequence Data Augmentation
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 641-645
Open AccessRethinking Sequential Relationships: Improving Sequential Recommenders with Inter-Sequence Data Augmentation
- Open Access
- 30 April 2023
- Published by Association for Computing Machinery (ACM)
- p. 864-868
Improving Product Search with Season-Aware Query-Product Semantic Similarity
Product search for online shopping should be season-aware, i.e., presenting seasonally relevant products to customers. In this paper, we propose a simple yet effective solution to improve seasonal relevance in product search by incorporating seasonality into language models for semantic matching. We first identify seasonal queries and products by analyzing implicit seasonal contexts through time-series analysis over the past year. Then we introduce explicit seasonal contexts by enhancing the query representation with a season token according to when the query is issued. A new season-enhanced BERT model (SE-BERT) is also proposed to learn the semantic similarity between the resulting seasonal queries and products. SE-BERT utilizes Multi-modal Adaption Gate (MAG) to augment the season-enhanced semantic embedding with other contextual information such as product price and review counts for robust relevance prediction. To better align with the ranking objective, a listwise loss function (neural NDCG) is used to regularize learning. Experimental results validate the effectiveness of the proposed method, which outperforms existing solutions for query-product relevance prediction in terms of NDCG and Price Weighted Purchases (PWP).
- 30 April 2023
- Published by Association for Computing Machinery (ACM)
- p. 864-868
Open AccessImproving Product Search with Season-Aware Query-Product Semantic Similarity
Product search for online shopping should be season-aware, i.e., presenting seasonally relevant products to customers. In this paper, we propose a simple yet effective solution to improve seasonal relevance in product search by incorporating seasonality into language models for semantic matching. We first identify seasonal queries and products by analyzing implicit seasonal contexts through time-series analysis over the past year. Then we introduce explicit seasonal contexts by enhancing the query representation with a season token according to when the query is issued. A new season-enhanced BERT model (SE-BERT) is also proposed to learn the semantic similarity between the resulting seasonal queries and products. SE-BERT utilizes Multi-modal Adaption Gate (MAG) to augment the season-enhanced semantic embedding with other contextual information such as product price and review counts for robust relevance prediction. To better align with the ranking objective, a listwise loss function (neural NDCG) is used to regularize learning. Experimental results validate the effectiveness of the proposed method, which outperforms existing solutions for query-product relevance prediction in terms of NDCG and Price Weighted Purchases (PWP).
- Open Access
- 14 November 2022
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 39 (1)
Digitally predicting protein localization and manipulating protein activity in fluorescence images using 4D reslicing GAN
While multi-channel fluorescence microscopy is a vital imaging method in biological studies, the number of channels that can be imaged simultaneously is limited by technical and hardware limitations such as emission spectra cross-talk. One solution is using deep neural networks to model the localization relationship between two proteins so that the localization of one protein can be digitally predicted. Furthermore, the input and predicted localization implicitly reflect the modeled relationship. Accordingly, observing the response of the prediction via manipulating input localization could provide an informative way to analyze the modeled relationships between the input and the predicted proteins. We propose a protein localization prediction (PLP) method using a cGAN named 4D Reslicing Generative Adversarial Network (4DR-GAN) to digitally generate additional channels. 4DR-GAN models the joint probability distribution of input and output proteins by simultaneously incorporating the protein localization signals in four dimensions including space and time. Because protein localization often correlates with protein activation state, based on accurate PLP, we further propose two novel tools: digital activation (DA) and digital inactivation (DI) to digitally activate and inactivate a protein, in order to observing the response of the predicted protein localization. Compared with genetic approaches, these tools allow precise spatial and temporal control. A comprehensive experiment on six pairs of proteins shows that 4DR-GAN achieves higher-quality PLP than Pix2Pix, and the DA and DI responses are consistent with the known protein functions. The proposed PLP method helps simultaneously visualize additional proteins, and the developed DA and DI tools provide guidance to study localization-based protein functions. The open-source code is available at https://github.com/YangJiaoUSA/4DR-GAN. Supplementary data are available at Bioinformatics online.
- 14 November 2022
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 39 (1)
Open AccessDigitally predicting protein localization and manipulating protein activity in fluorescence images using 4D reslicing GAN
While multi-channel fluorescence microscopy is a vital imaging method in biological studies, the number of channels that can be imaged simultaneously is limited by technical and hardware limitations such as emission spectra cross-talk. One solution is using deep neural networks to model the localization relationship between two proteins so that the localization of one protein can be digitally predicted. Furthermore, the input and predicted localization implicitly reflect the modeled relationship. Accordingly, observing the response of the prediction via manipulating input localization could provide an informative way to analyze the modeled relationships between the input and the predicted proteins. We propose a protein localization prediction (PLP) method using a cGAN named 4D Reslicing Generative Adversarial Network (4DR-GAN) to digitally generate additional channels. 4DR-GAN models the joint probability distribution of input and output proteins by simultaneously incorporating the protein localization signals in four dimensions including space and time. Because protein localization often correlates with protein activation state, based on accurate PLP, we further propose two novel tools: digital activation (DA) and digital inactivation (DI) to digitally activate and inactivate a protein, in order to observing the response of the predicted protein localization. Compared with genetic approaches, these tools allow precise spatial and temporal control. A comprehensive experiment on six pairs of proteins shows that 4DR-GAN achieves higher-quality PLP than Pix2Pix, and the DA and DI responses are consistent with the known protein functions. The proposed PLP method helps simultaneously visualize additional proteins, and the developed DA and DI tools provide guidance to study localization-based protein functions. The open-source code is available at https://github.com/YangJiaoUSA/4DR-GAN. Supplementary data are available at Bioinformatics online.
- 30 October 2022
- Published by Springer Nature
- p. 19-35
Fine-Grained Fashion Representation Learning by Online Deep Clustering
- 30 October 2022
- Published by Springer Nature
- p. 19-35
Fine-Grained Fashion Representation Learning by Online Deep Clustering
- Open Access
Digitally Predicting Protein Localization and Manipulating Protein Activity in Fluorescence Images Using Four-dimensional Reslicing GAN
Motivation While multi-channel fluorescence microscopy is a vital imaging method in biological studies, the number of channels that can be imaged simultaneously is limited by technical and hardware limitations such as emission spectra cross-talk. One feasible solution is using deep neural networks to model the localization relationship between two proteins so that the localization of a protein can be digitally predicted. Furthermore, the input and predicted localization implicitly reflects the modeled relationship. Accordingly, observing the predictions via repeatedly manipulating input localizations is an explainable and feasible way to analyze the modeled relationships between the input and the predicted proteins. Results We propose a Protein Localization Prediction (PLP) method using a cGAN named Four-dimensional Reslicing Generative Adversarial Network (4DR-GAN) to digitally generate additional channels. 4DR-GAN models the joint probability distribution of imaged and target proteins by simultaneously incorporating the protein localization signals in four dimensions including space and time. Because protein localization often correlates with protein activation state, with accurate PLP, we further propose two novel tools: digital activation (DA) and digital inactivation (DI) to digitally activate and inactivate a protein and observe the response of the predicted protein localization. Compared with genetic approaches, these tools allow precise spatial and temporal control. A comprehensive experiment on four groups of proteins shows that 4DR-GAN achieves higher-quality PLP than Pix2Pix and the DA and DI responses are consistent with the known protein functions. The proposed PLP method helps simultaneously visualize additional proteins and DA and DI provide guidance to study localization-based protein functions. Availability and Implementation The open-source code is at https://github.com/YangJiaoUSA/4DR-GAN.
Open AccessDigitally Predicting Protein Localization and Manipulating Protein Activity in Fluorescence Images Using Four-dimensional Reslicing GAN
Motivation While multi-channel fluorescence microscopy is a vital imaging method in biological studies, the number of channels that can be imaged simultaneously is limited by technical and hardware limitations such as emission spectra cross-talk. One feasible solution is using deep neural networks to model the localization relationship between two proteins so that the localization of a protein can be digitally predicted. Furthermore, the input and predicted localization implicitly reflects the modeled relationship. Accordingly, observing the predictions via repeatedly manipulating input localizations is an explainable and feasible way to analyze the modeled relationships between the input and the predicted proteins. Results We propose a Protein Localization Prediction (PLP) method using a cGAN named Four-dimensional Reslicing Generative Adversarial Network (4DR-GAN) to digitally generate additional channels. 4DR-GAN models the joint probability distribution of imaged and target proteins by simultaneously incorporating the protein localization signals in four dimensions including space and time. Because protein localization often correlates with protein activation state, with accurate PLP, we further propose two novel tools: digital activation (DA) and digital inactivation (DI) to digitally activate and inactivate a protein and observe the response of the predicted protein localization. Compared with genetic approaches, these tools allow precise spatial and temporal control. A comprehensive experiment on four groups of proteins shows that 4DR-GAN achieves higher-quality PLP than Pix2Pix and the DA and DI responses are consistent with the known protein functions. The proposed PLP method helps simultaneously visualize additional proteins and DA and DI provide guidance to study localization-based protein functions. Availability and Implementation The open-source code is at https://github.com/YangJiaoUSA/4DR-GAN.
- 9 April 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Conference on Computer Vision and Pattern Recognition (CVPR)
- Vol. 2019, 1087-1096
Multi-Object Portion Tracking in 4D Fluorescence Microscopy Imagery With Deep Feature Maps
3D fluorescence microscopy of living organisms has increasingly become an essential and powerful tool in biomedical research and diagnosis. Exploding amount of imaging data has been collected whereas efficient and effective computational tools to extract information from them are still lagged behind. This largely is due to the challenges in analyzing biological data. Interesting biological structures are not only small but often are morphologically irregular and highly dynamic. Tracking cells in live organisms has been studied for years as a sophisticated mission in bioinformatics. However, existing tracking methods for cells are not effective in tracking subcellular structures, such as protein complexes, which feature in continuous morphological changes, such as split and merge, in addition to fast migration and complex motion. In this paper, we first define the problem of multi-object portion tracking to model protein object tracking process. A multi-object tracking method with portion matching is proposed based on 3D segmentation results. The proposed method distills deep feature maps from deep networks, then recognizes and matches objects' portions using extended search. Experimental results confirm that the proposed method achieves 2.98% higher on consistent tracking accuracy and 35.48% higher on event identification accuracy.
- 9 April 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Conference on Computer Vision and Pattern Recognition (CVPR)
- Vol. 2019, 1087-1096
Multi-Object Portion Tracking in 4D Fluorescence Microscopy Imagery With Deep Feature Maps
3D fluorescence microscopy of living organisms has increasingly become an essential and powerful tool in biomedical research and diagnosis. Exploding amount of imaging data has been collected whereas efficient and effective computational tools to extract information from them are still lagged behind. This largely is due to the challenges in analyzing biological data. Interesting biological structures are not only small but often are morphologically irregular and highly dynamic. Tracking cells in live organisms has been studied for years as a sophisticated mission in bioinformatics. However, existing tracking methods for cells are not effective in tracking subcellular structures, such as protein complexes, which feature in continuous morphological changes, such as split and merge, in addition to fast migration and complex motion. In this paper, we first define the problem of multi-object portion tracking to model protein object tracking process. A multi-object tracking method with portion matching is proposed based on 3D segmentation results. The proposed method distills deep feature maps from deep networks, then recognizes and matches objects' portions using extended search. Experimental results confirm that the proposed method achieves 2.98% higher on consistent tracking accuracy and 35.48% higher on event identification accuracy.
- 20 May 2019
- Published by SPIE-Intl Soc Optical Eng in Journal of Medical Imaging
- Vol. 6 (02) , 024006-13
DeepQuantify: deep learning and quantification system of white blood cells in light microscopy images of injured skeletal muscles
White blood cells (WBCs) are the most diverse types of cells observed in the healing process of injured skeletal muscles. In the recovery process, WBCs exhibit a dynamic cellular response and undergo multiple changes of the protein expression. The progress of healing can be analyzed by the number of WBCs or by the number of specific proteins observed in light microscopy images obtained at different time points after injury. We propose a deep learning quantification and analysis system called DeepQuantify to analyze WBCs in light microscopy images of uninjured and injured muscles of female mice. The DeepQuantify system features in segmentation using the localized iterative Otsu’s thresholding method, masking postprocessing, and classification of WBCs with a convolutional neural network (CNN) classifier to achieve a high accuracy and a low manpower cost. The proposed two-layer CNN classifier designed based on the optimization hypothesis is evaluated and compared with other CNN classifiers. The DeepQuantify system adopting these CNN classifiers is evaluated for quantifying CD68-positive macrophages and 7/4-positive neutrophils and compared with the state-of-the-art deep learning segmentation architectures. DeepQuantify achieves an accuracy of 90.64% and 89.31% for CD68-positive macrophages and 7/4-positive neutrophils, respectively. The DeepQuantify system employing the proposed two-layer CNN architecture achieves better performance than those deep segmentation architectures. The quantitative analysis of two protein dynamics during muscle recovery is also presented.
- 20 May 2019
- Published by SPIE-Intl Soc Optical Eng in Journal of Medical Imaging
- Vol. 6 (02) , 024006-13
DeepQuantify: deep learning and quantification system of white blood cells in light microscopy images of injured skeletal muscles
White blood cells (WBCs) are the most diverse types of cells observed in the healing process of injured skeletal muscles. In the recovery process, WBCs exhibit a dynamic cellular response and undergo multiple changes of the protein expression. The progress of healing can be analyzed by the number of WBCs or by the number of specific proteins observed in light microscopy images obtained at different time points after injury. We propose a deep learning quantification and analysis system called DeepQuantify to analyze WBCs in light microscopy images of uninjured and injured muscles of female mice. The DeepQuantify system features in segmentation using the localized iterative Otsu’s thresholding method, masking postprocessing, and classification of WBCs with a convolutional neural network (CNN) classifier to achieve a high accuracy and a low manpower cost. The proposed two-layer CNN classifier designed based on the optimization hypothesis is evaluated and compared with other CNN classifiers. The DeepQuantify system adopting these CNN classifiers is evaluated for quantifying CD68-positive macrophages and 7/4-positive neutrophils and compared with the state-of-the-art deep learning segmentation architectures. DeepQuantify achieves an accuracy of 90.64% and 89.31% for CD68-positive macrophages and 7/4-positive neutrophils, respectively. The DeepQuantify system employing the proposed two-layer CNN architecture achieves better performance than those deep segmentation architectures. The quantitative analysis of two protein dynamics during muscle recovery is also presented.
- 10 November 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 52 (4) , 2030-2038
Novel suboptimal filter via higher order central moments
In this paper, we construct a new suboptimal filter by deriving the Ito's stochastic differential equations of the estimation of higher order central moments, satisfy, and impose some conditions to form a closed system. The essentially infinite-dimensional cubic sensor problem has been investigated in detail numerically to illustrate the reasonableness of the imposed conditions, and the numerical experiments support our discussion. A two-dimensional polynomial filtering problem has also been experimented.
- 10 November 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 52 (4) , 2030-2038
Novel suboptimal filter via higher order central moments
In this paper, we construct a new suboptimal filter by deriving the Ito's stochastic differential equations of the estimation of higher order central moments, satisfy, and impose some conditions to form a closed system. The essentially infinite-dimensional cubic sensor problem has been investigated in detail numerically to illustrate the reasonableness of the imposed conditions, and the numerical experiments support our discussion. A two-dimensional polynomial filtering problem has also been experimented.