Machine Learning at Microsoft with ML.NET

Ahmed, Zeeshan; Amizadeh, Saeed; Bilenko, Mikhail; Carr, Robert D.; Chin, Wei-Sheng; Dekel, Yael; Dupre, Xavier; Eksarevskiy, Vadim; Filipi, Senja; Finley, Tom; Goswami, Abhishek; Hoover, Monte; Inglis, Scott; Interlandi, Matteo; Kazmi, Najeeb; Krivosheev, Gleb; Luferenko, Pete; Matantsev, Ivan; Matusevych, Sergiy; Moradi, Shahab; Nazirov, Gani; Ormont, Justin; Oshri, Gal; Pagnoni, Artidoro; Parmar, Jignesh; Roy, Prabhat; Siddiqui, Mohammad Zeeshan; Weimer, Markus; Zahirazami, Shauheen; Zhu, Yiwen

doi:10.1145/3292500.3330667

Cited by 47 publications

(35 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Several recent efforts aim to simplify ML development through a general-purpose machine learning system with both training and serving of models [2,5,6,13,15,26,34,35,42,68,69]. Some of these systems that share similar goals of MLIoT are the end-toend "ML Platforms" that run at commercial settings.…”

Section: Training/serving Hybrid Systemsmentioning

confidence: 99%

“…Such systems generally run on the cloud incurring a higher cost for better workload environments and restrict users to a specific set of algorithms or libraries, so users are on their own when they step outside these boundaries. Similarly, Other commercial IoT tailored systems such as Google's Cloud IoT [26], Microsoft's ML.net [2] and AWS IoT Greengrass [5] are in-house proprietary systems focused on specific hardware and ML algorithms. Academic approaches such as Velox [13] and InferLine [15] propose managing the lifecycle of model training, serving, and updating but they are intended towards modeling efficient execution of ML pipelines to reduce cost or meet the SLO constraints.…”

Section: Training/serving Hybrid Systemsmentioning

confidence: 99%

See 1 more Smart Citation

MLIoT

Boovaraghavan

Maravi

Mallela

et al. 2021

Proceedings of the International Conference on Internet-of-Things Design and Implementation

View full text Add to dashboard Cite

Modern Internet of Things (IoT) applications, from contextual sensing to voice assistants, rely on ML-based training and serving systems using pre-trained models to render predictions. However, real-world IoT environments are diverse, with rich IoT sensors and need ML models to be personalized for each setting using relatively less training data. Most existing general-purpose ML systems are optimized for specific and dedicated hardware resources and do not adapt to changing resources and different IoT application requirements. To address this gap, we propose MLIoT, an end-to-end Machine Learning System tailored towards supporting the entire lifecycle of IoT applications. MLIoT adapts to different IoT data sources, IoT tasks, and compute resources by automatically training, optimizing, and serving models based on expressive applicationspecific policies. MLIoT also adapts to changes in IoT environments or compute resources by enabling re-training, and updating models served on the fly while maintaining accuracy and performance. Our evaluation across a set of benchmarks show that MLIoT can handle multiple IoT tasks, each with individual requirements, in a scalable manner while maintaining high accuracy and performance. We compare MLIoT with two state-of-the-art hand-tuned systems and a commercial ML system showing that MLIoT improves accuracy from 50% -75% while reducing or maintaining latency. CCS CONCEPTS• Computer systems organization → Client-server architectures; • Computing methodologies → Artificial intelligence.

show abstract

Section: Training/serving Hybrid Systemsmentioning

confidence: 99%

Section: Training/serving Hybrid Systemsmentioning

confidence: 99%

MLIoT

Boovaraghavan

Maravi

Mallela

et al. 2021

Proceedings of the International Conference on Internet-of-Things Design and Implementation

View full text Add to dashboard Cite

show abstract

“…Kolejne powstające platformy mają zminimalizować problemy z wydajnością, unifikacją, rozszerzalnością i skalowalnością, umożliwiając jak najlepszą wydajność przy maksymalnym wykorzystaniu zasobów sprzętowych [13].…”

Section: Wstępunclassified

Performance analysis of machine learning libraries

Kędziora¹,

Maksim²

2021

jcsi

View full text Add to dashboard Cite

The paper presents results of performance analysis of machine learning libraries. The research was based on ML.NET and TensorFlow tools. The analysis was based on a comparison of running time of the libraries, during detection of objects on sets of images, using hardware with different parameters. The library, consuming fewer hardware resources, turned out to be TensorFlow. The choice of hardware platform and the possibility of using graphic cores, affecting the increase in computational efficiency, turned out to be not without significance.

show abstract

“…In the first iteration, we built a simple multiple additive regression tree (MART) and partition into multiple models to scale up with the amount of data and classes (teams) in IMS. We build a one-vs-all FastTree [1] binary classifier, a MART using gradient boosting algorithm in ML.NET. In each training step, FastTree builds a decision tree to add to the ensemble of trees from previous steps.…”

Section: Multiple Additive Regressionmentioning

confidence: 99%

DeepTriage

Pham

Jain

Dauterman

et al. 2020

Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery &Amp; Data Mining

Self Cite

View full text Add to dashboard Cite

As cloud services are growing and generating high revenues, the cost of downtime in these services is becoming significantly expensive. To reduce loss and service downtime, a critical primary step is to execute incident triage, the process of assigning a service incident to the correct responsible team, in a timely manner. An incorrect assignment risks additional incident reroutings and increases its time to mitigate by 10x. However, automated incident triage in large cloud services faces many challenges: (1) a highly imbalanced incident distribution from a large number of teams, (2) wide variety in formats of input data or data sources, (3) scaling to meet production-grade requirements, and (4) gaining engineers' trust in using machine learning recommendations. To address these challenges, we introduce DeepTriage, an intelligent incident transfer service combining multiple machine learning techniques-gradient boosted classifiers, clustering methods, and deep neural networks-in an ensemble to recommend the responsible team to triage an incident. Experimental results on real incidents in Microsoft Azure show that our service achieves 82.9% F1 score. For highly impacted incidents, DeepTriage achieves F1 score from 76.3% − 91.3%. We have applied best practices and state-of-the-art frameworks to scale DeepTriage to handle incident routing for all cloud services. Deep-Triage has been deployed in Azure since October 2017 and is used by thousands of teams daily. CCS CONCEPTS • Computing methodologies → Machine learning; Natural language processing; • Applied computing → Business process management.

show abstract

Machine Learning at Microsoft with ML.NET

Cited by 47 publications

References 10 publications

MLIoT

MLIoT

Performance analysis of machine learning libraries

DeepTriage

Contact Info

Product

Resources

About