arXiv subjects

Diksha Garg

Publications and source records attributed to Diksha Garg.

21 records · Page 2Linked to original sources

arXiv2021

NISER: Normalized Item and Session Representations to Handle Popularity Bias

The goal of session-based recommendation (SR) models is to utilize the information from past actions (e.g. item/product clicks) in a session to recommend items that a user is likely to click next. Recently it has been shown that the sequence of item interactions in a session can be modeled as graph-structured data to better account for complex item transitions. Graph neural networks (GNNs) can learn useful representations for such session-graphs, and have been shown to improve over sequential models such as recurrent neural networks [14]. However, we note that these GNN-based recommendation models suffer from popularity bias: the models are biased towards recommending popular items, and fail to recommend relevant long-tail items (less popular or less frequent items). Therefore, these models perform poorly for the less popular new items arriving daily in a practical online setting. We demonstrate that this issue is, in part, related to the magnitude or norm of the learned item and session-graph representations (embedding vectors). We propose a training procedure that mitigates this issue by using normalized representations. The models using normalized item and session-graph representations perform significantly better: i. for the less popular long-tail items in the offline setting, and ii. for the less popular newly introduced items in the online setting. Furthermore, our approach significantly improves upon existing state-of-the-art on three benchmark datasets.

cs.IR↗

arXiv2020

Batch-Constrained Distributional Reinforcement Learning for Session-based Recommendation

Most of the existing deep reinforcement learning (RL) approaches for session-based recommendations either rely on costly online interactions with real users, or rely on potentially biased rule-based or data-driven user-behavior models for learning. In this work, we instead focus on learning recommendation policies in the pure batch or offline setting, i.e. learning policies solely from offline historical interaction logs or batch data generated from an unknown and sub-optimal behavior policy, without further access to data from the real-world or user-behavior models. We propose BCD4Rec: Batch-Constrained Distributional RL for Session-based Recommendations. BCD4Rec builds upon the recent advances in batch (offline) RL and distributional RL to learn from offline logs while dealing with the intrinsically stochastic nature of rewards from the users due to varied latent interest preferences (environments). We demonstrate that BCD4Rec significantly improves upon the behavior policy as well as strong RL and non-RL baselines in the batch setting in terms of standard performance metrics like Click Through Rates or Buy Rates. Other useful properties of BCD4Rec include: i. recommending items from the correct latent categories indicating better value estimates despite large action space (of the order of number of items), and ii. overcoming popularity bias in clicked or bought items typically present in the offline logs.

cs.LG↗

arXiv2020

The ABC130 barrel module prototyping programme for the ATLAS strip tracker

For the Phase-II Upgrade of the ATLAS Detector, its Inner Detector, consisting of silicon pixel, silicon strip and transition radiation sub-detectors, will be replaced with an all new 100 % silicon tracker, composed of a pixel tracker at inner radii and a strip tracker at outer radii. The future ATLAS strip tracker will include 11,000 silicon sensor modules in the central region (barrel) and 7,000 modules in the forward region (end-caps), which are foreseen to be constructed over a period of 3.5 years. The construction of each module consists of a series of assembly and quality control steps, which were engineered to be identical for all production sites. In order to develop the tooling and procedures for assembly and testing of these modules, two series of major prototyping programs were conducted: an early program using readout chips designed using a 250 nm fabrication process (ABCN-25) and a subsequent program using a follow-up chip set made using 130 nm processing (ABC130 and HCC130 chips). This second generation of readout chips was used for an extensive prototyping program that produced around 100 barrel-type modules and contributed significantly to the development of the final module layout. This paper gives an overview of the components used in ABC130 barrel modules, their assembly procedure and findings resulting from their tests.

physics.ins-det↗