Publications

2024

1. Rethinking ’Complement’ Recommendations at Scale with SIMD

   Shrey Pandey, Saikat Kumar Das , Hrishikesh V. Ganu , and 1 more author

   In Proceedings of the 15th ACM/SPEC International Conference on Performance Engineering , , London, United Kingdom, , 2024

   Abs Bib

   Maximizing cart value by increasing the number of items in electronic carts is one of the key strategies adopted by e-commerce platforms for optimal conversion of positive user intent during an online shopping session. Recommender systems play a key-role in suggesting personalized candidate items that can be added to cart by the user. However, it is important to serve a diverse set of personalized recommendations that ’complement’ user’s cart content to practically increase item count in cart and also contribute towards product discovery. Borrowed from Quantum Physics, Determinantal Point Processes (DPP) are used widely in recommender systems to diversify personalized product recommendations for improved user engagement. However, vertically scaling DPP for recommendation sets, personalized with vector similarity metric like cosine similarity, to serve large scale real-time concurrent user requests is non-trivial. We propose a vectorized reformulation of cosine similarity and conditional DPP implementation to best utilize the highly improved vector computation capabilities (SIMD) of modern processors. Experimental evidence on real-world traffic shows that the proposed method can handle upto 15x more concurrent traffic while improving latency. The proposed method also uses portable SIMD constructs from Python libraries which can be easily adopted in most available SIMD supported CPUs with minimal code changes.

   @inproceedings{10.1145/3629526.3645041,
     author = {Pandey, Shrey and Das, Saikat Kumar and Ganu, Hrishikesh V. and Singh, Satyajeet},
     title = {Rethinking 'Complement' Recommendations at Scale with SIMD},
     year = {2024},
     isbn = {9798400704444},
     publisher = {Association for Computing Machinery},
     address = {New York, NY, USA},
     url = {https://doi.org/10.1145/3629526.3645041},
     doi = {10.1145/3629526.3645041},
     booktitle = {Proceedings of the 15th ACM/SPEC International Conference on Performance Engineering},
     pages = {25–36},
     numpages = {12},
     keywords = {determinantal point processes, diversification, recommendations, recommender systems, simd, vectorization},
     location = {<conf-loc>, <city>London</city>, <country>United Kingdom</country>, </conf-loc>},
     series = {ICPE '24},
   }

2023

1. Efficient Vector Store System for Python using Shared Memory

   Dhruv Patel , Shrey Pandey, and Abhishek Sharma

   In Proceedings of the Second International Conference on AI-ML Systems , Bangalore, India, 2023

   Abs Bib

   Many e-commerce companies use machine learning to make customer experience better. Even within a single company, there will be generally many independent services running, each specializing in some aspect of customer experience. Since machine learning models work on abstract vectors representing users and/or items, each such service needs a way to store these vectors. A common approach nowadays is to save them in in-memory caches like Memcached. As these caches run in their own processes, and Machine Learning services generally run as Python services, there is a communication overhead involved for each request that ML service serves. One can reduce this overhead by directly storing these vectors in a Python dictionary within the service. To support concurrency and scale, a single node runs multiple instances of the same service. Thus, we also want to avoid duplicating these vectors across multiple processes. In this paper, we propose a system to store vectors in shared memory and efficiently serve all concurrent instances of the service, without replicating the vectors themselves. We achieve up to 4.5x improvements in latency compared to Memcached. Additionally, due to availability of more memory, we can increase the number of server processes running in each node, translating into greater throughput. We also discuss the impact of the proposed method (towards increasing the throughput) in live production scenario.

   @inproceedings{10.1145/3564121.3564799,
     author = {Patel, Dhruv and Pandey, Shrey and Sharma, Abhishek},
     title = {Efficient Vector Store System for Python using Shared Memory},
     year = {2023},
     isbn = {9781450398473},
     publisher = {Association for Computing Machinery},
     address = {New York, NY, USA},
     url = {https://doi.org/10.1145/3564121.3564799},
     doi = {10.1145/3564121.3564799},
     booktitle = {Proceedings of the Second International Conference on AI-ML Systems},
     articleno = {22},
     numpages = {6},
     keywords = {vector data store, shared memory, embedding data store},
     location = {Bangalore, India},
     series = {AIMLSystems '22},
   }

2021

1. CLOTON: A GAN based approach for Clothing Try-On

   Shrey Pandey, Yash Srivastava , Yukta Meena , and 1 more author

   In 2021 8th International Conference on Signal Processing and Integrated Networks (SPIN) , Aug 2021

   Abs Bib

   Cloton is a framework that produces a picture of a person wearing the target clothing while taking into account arbitrary body shapes and poses. As the generation of an individual image necessitates the extraction of body shape and pose from one image and target clothing from another, the architecture is divided into a number of tasks. With these two images as input, the system generates a new picture of the user wearing the target clothing using a generative network. The output is this generated image, which is then sent back to the user. By using encoder-decoder networks as the generator model and training it using the generative adversarial approach used by GANs, Cloton manages to generate visually appealing garment swapped images of the user. Cloton also has good results in the quantitative analysis of the system as compared to previous works.

   @inproceedings{9565973,
     author = {Pandey, Shrey and Srivastava, Yash and Meena, Yukta and Dewang, Rupesh Kumar},
     booktitle = {2021 8th International Conference on Signal Processing and Integrated Networks (SPIN)},
     title = {CLOTON: A GAN based approach for Clothing Try-On},
     year = {2021},
     volume = {},
     number = {},
     pages = {595-601},
     keywords = {Training;Shape;Statistical analysis;Clothing;Fitting;Signal processing;Generative adversarial networks},
     doi = {10.1109/SPIN52536.2021.9565973},
     issn = {2688-769X},
     month = aug
   }