Learning-to-rank (LTR) models rank items based on specific features, aiming to maximize ranking utility by prioritizing highly relevant items. However, optimizing only for ranking utility can lead to representational harm and may fail to address implicit bias in relevance scores. Prior studies introduced algorithms to…
Recent research on time-series self-supervised models shows great promise in learning semantic representations. However, it has been limited to small-scale datasets, e.g., thousands of temporal sequences. In this work, we make key technical contributions that are tailored to the numerical properties of time-series data and…
Learning-to-rank (LTR) models rank items based on specific features, aiming to maximize ranking utility by prioritizing highly relevant items. However, optimizing only for ranking utility can lead to representational harm and may fail to address implicit bias in relevance scores. Prior studies introduced algorithms to…
Learning-to-rank (LTR) models rank items based on specific features, aiming to maximize ranking utility by prioritizing highly relevant items. However, optimizing only for ranking utility can lead to representational harm and may fail to address implicit bias in relevance scores. Prior studies introduced algorithms to…
In large-scale cloud service systems, monitoring metric data and conducting anomaly detection is an important way to maintain reliability and stability. However, great disparity exists between academic approaches and industrial practice to anomaly detection. Industry predominantly uses simple, efficient methods due to better interpretability and…
Serving large language models (LLMs) in production can incur substantial costs, which has prompted recent advances in inference system optimizations. Today, these systems are evaluated against conventional latency and throughput metrics (eg. TTFT, TBT, Normalised Latency and TPOT). However, these metrics fail to fully capture…
In-context learning for vision data has been underexplored compared with that in natural language. Previous works studied image in-context learning, urging models to generate a single image guided by demonstrations. In this paper, we propose and study video in-context learning, where the model starts from…
Controlling the evolution of complex physical systems is a fundamental task across science and engineering. Classical techniques suffer from limited applicability or huge computational costs. On the other hand, recent deep learning and reinforcement learning-based approaches often struggle to optimize long-term control sequences under the…
Major cloud providers have committed to lowering carbon emissions by 2030 across their datacenters, and research has contributed many ideas on how this may be achieved. However, a major contributor to datacenter emissions has not received enough attention: storage. Storage - everything from file storage…
On July 5th, 2024, Microsoft Research Asia (MSRA) hosted the MSRA Day as part of the Summer Institute in Computational Social Science (SICSS). This event, held at MSRA’s headquarters, featured a series of impactful presentations and discussions on the intersection of artificial intelligence (AI) and…
