�ݺ�ߣshows by User: SigOpt

�ݺ�ߣshows by User: SigOpt / http://www.slideshare.net/images/logo.gif �ݺ�ߣshows by User: SigOpt / Thu, 20 Aug 2020 22:01:09 GMT �ݺ�ߣShare feed for �ݺ�ߣshows by User: SigOpt Optimizing BERT and Natural Language Models with SigOpt Experiment Management /slideshow/optimizing-bert-and-natural-language-models-with-sigopt-experiment-management/238084938 multimetricbertf9webinar-200820220109
SigOpt Machine Learning Engineer Meghana Ravikumar explains how she reduced the size of a BERT natural language model trained on the SQUAD 2.0 question-answer database, to reduce its size while maintaining performance using a "distillation" process optimized with SigOpt's Experiment Management functionality.]]>
SigOpt Machine Learning Engineer Meghana Ravikumar explains how she reduced the size of a BERT natural language model trained on the SQUAD 2.0 question-answer database, to reduce its size while maintaining performance using a "distillation" process optimized with SigOpt's Experiment Management functionality.]]> Thu, 20 Aug 2020 22:01:09 GMT /slideshow/optimizing-bert-and-natural-language-models-with-sigopt-experiment-management/238084938 SigOpt@slideshare.net(SigOpt) Optimizing BERT and Natural Language Models with SigOpt Experiment Management SigOpt SigOpt Machine Learning Engineer Meghana Ravikumar explains how she reduced the size of a BERT natural language model trained on the SQUAD 2.0 question-answer database, to reduce its size while maintaining performance using a "distillation" process optimized with SigOpt's Experiment Management functionality. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/multimetricbertf9webinar-200820220109-thumbnail.jpg?width=120&height=120&fit=bounds" /> SigOpt Machine Learning Engineer Meghana Ravikumar explains how she reduced the size of a BERT natural language model trained on the SQUAD 2.0 question-answer database, to reduce its size while maintaining performance using a "distillation" process optimized with SigOpt's Experiment Management functionality.

Optimizing BERT and Natural Language Models with SigOpt Experiment Management from SigOpt

]]> 228 1 https://cdn.slidesharecdn.com/ss_thumbnails/multimetricbertf9webinar-200820220109-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Experiment Management for the Enterprise /slideshow/experiment-management-for-the-enterprise/236800355 f9webinar-sigopt-mlopssolutionfortheenterprise-200710232253
SigOpt's Fay Kallel, Head of Product, and Jim Blomo, Head of Engineering, describe the latest updates to SigOpt, a suite of features that help you manage your modeling process.]]>
SigOpt's Fay Kallel, Head of Product, and Jim Blomo, Head of Engineering, describe the latest updates to SigOpt, a suite of features that help you manage your modeling process.]]> Fri, 10 Jul 2020 23:22:52 GMT /slideshow/experiment-management-for-the-enterprise/236800355 SigOpt@slideshare.net(SigOpt) Experiment Management for the Enterprise SigOpt SigOpt's Fay Kallel, Head of Product, and Jim Blomo, Head of Engineering, describe the latest updates to SigOpt, a suite of features that help you manage your modeling process. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/f9webinar-sigopt-mlopssolutionfortheenterprise-200710232253-thumbnail.jpg?width=120&height=120&fit=bounds" /> SigOpt's Fay Kallel, Head of Product, and Jim Blomo, Head of Engineering, describe the latest updates to SigOpt, a suite of features that help you manage your modeling process.

Experiment Management for the Enterprise from SigOpt

]]> 187 0 https://cdn.slidesharecdn.com/ss_thumbnails/f9webinar-sigopt-mlopssolutionfortheenterprise-200710232253-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Efficient NLP by Distilling BERT and Multimetric Optimization /slideshow/efficient-nlp-by-distilling-bert-and-multimetric-optimization/236223976 webinarmultimetricbert-200625234510
SigOpt ML Engineer Meghana Ravikumar explains how to use multimetric optimization to achieve a more efficient, compact BERT model to perform on a question-answering task.]]>
SigOpt ML Engineer Meghana Ravikumar explains how to use multimetric optimization to achieve a more efficient, compact BERT model to perform on a question-answering task.]]> Thu, 25 Jun 2020 23:45:10 GMT /slideshow/efficient-nlp-by-distilling-bert-and-multimetric-optimization/236223976 SigOpt@slideshare.net(SigOpt) Efficient NLP by Distilling BERT and Multimetric Optimization SigOpt SigOpt ML Engineer Meghana Ravikumar explains how to use multimetric optimization to achieve a more efficient, compact BERT model to perform on a question-answering task. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/webinarmultimetricbert-200625234510-thumbnail.jpg?width=120&height=120&fit=bounds" /> SigOpt ML Engineer Meghana Ravikumar explains how to use multimetric optimization to achieve a more efficient, compact BERT model to perform on a question-answering task.

Efficient NLP by Distilling BERT and Multimetric Optimization from SigOpt

]]> 279 0 https://cdn.slidesharecdn.com/ss_thumbnails/webinarmultimetricbert-200625234510-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Detecting COVID-19 Cases with Deep Learning /slideshow/detecting-covid19-cases-with-deep-learning/235359561 covid-netwebinar-200610194247
SigOpt Research Engineer Michael McCourt and DarwinAI CTO Alexander Wong explain how they used SigOpt and hyperparameter optimization to successfully improve accuracy of detecting COVID-19 cases from chest X-Rays, using the COVID-Net model and the COVIDx open dataset.]]>
SigOpt Research Engineer Michael McCourt and DarwinAI CTO Alexander Wong explain how they used SigOpt and hyperparameter optimization to successfully improve accuracy of detecting COVID-19 cases from chest X-Rays, using the COVID-Net model and the COVIDx open dataset.]]> Wed, 10 Jun 2020 19:42:46 GMT /slideshow/detecting-covid19-cases-with-deep-learning/235359561 SigOpt@slideshare.net(SigOpt) Detecting COVID-19 Cases with Deep Learning SigOpt SigOpt Research Engineer Michael McCourt and DarwinAI CTO Alexander Wong explain how they used SigOpt and hyperparameter optimization to successfully improve accuracy of detecting COVID-19 cases from chest X-Rays, using the COVID-Net model and the COVIDx open dataset. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/covid-netwebinar-200610194247-thumbnail.jpg?width=120&height=120&fit=bounds" /> SigOpt Research Engineer Michael McCourt and DarwinAI CTO Alexander Wong explain how they used SigOpt and hyperparameter optimization to successfully improve accuracy of detecting COVID-19 cases from chest X-Rays, using the COVID-Net model and the COVIDx open dataset.

Detecting COVID-19 Cases with Deep Learning from SigOpt

]]> 185 0 https://cdn.slidesharecdn.com/ss_thumbnails/covid-netwebinar-200610194247-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Metric Management: a SigOpt Applied Use Case /slideshow/metric-management-a-sigopt-applied-use-case/234704593 metricmanagementdemocastwebinarslides-200528235644
These slides correspond to a recording of a live webcast of a demo of Metric Management functionality in SigOpt, keeping model size down while increasing validation accuracy for a road sign image classification problem.]]>
These slides correspond to a recording of a live webcast of a demo of Metric Management functionality in SigOpt, keeping model size down while increasing validation accuracy for a road sign image classification problem.]]> Thu, 28 May 2020 23:56:44 GMT /slideshow/metric-management-a-sigopt-applied-use-case/234704593 SigOpt@slideshare.net(SigOpt) Metric Management: a SigOpt Applied Use Case SigOpt These slides correspond to a recording of a live webcast of a demo of Metric Management functionality in SigOpt, keeping model size down while increasing validation accuracy for a road sign image classification problem. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/metricmanagementdemocastwebinarslides-200528235644-thumbnail.jpg?width=120&height=120&fit=bounds" /> These slides correspond to a recording of a live webcast of a demo of Metric Management functionality in SigOpt, keeping model size down while increasing validation accuracy for a road sign image classification problem.

Metric Management: a SigOpt Applied Use Case from SigOpt

]]> 183 0 https://cdn.slidesharecdn.com/ss_thumbnails/metricmanagementdemocastwebinarslides-200528235644-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Tuning for Systematic Trading: Talk 3: Training, Tuning, and Metric Strategy /slideshow/tuning-for-systematic-trading-talk-3-training-tuning-and-metric-strategy/234297131 talkingseries-presentation1-200519230409
This talk explains how you can train and tune efficiently using metric strategy to assign, store, and optimize a variety of metrics, even changing them over time. Tobias Andreassen, who supports a number of our systematic trading customers, explained how he helps customers tune more efficiently with these SigOpt features in real-world scenarios.]]>
This talk explains how you can train and tune efficiently using metric strategy to assign, store, and optimize a variety of metrics, even changing them over time. Tobias Andreassen, who supports a number of our systematic trading customers, explained how he helps customers tune more efficiently with these SigOpt features in real-world scenarios.]]> Tue, 19 May 2020 23:04:09 GMT /slideshow/tuning-for-systematic-trading-talk-3-training-tuning-and-metric-strategy/234297131 SigOpt@slideshare.net(SigOpt) Tuning for Systematic Trading: Talk 3: Training, Tuning, and Metric Strategy SigOpt This talk explains how you can train and tune efficiently using metric strategy to assign, store, and optimize a variety of metrics, even changing them over time. Tobias Andreassen, who supports a number of our systematic trading customers, explained how he helps customers tune more efficiently with these SigOpt features in real-world scenarios. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/talkingseries-presentation1-200519230409-thumbnail.jpg?width=120&height=120&fit=bounds" /> This talk explains how you can train and tune efficiently using metric strategy to assign, store, and optimize a variety of metrics, even changing them over time. Tobias Andreassen, who supports a number of our systematic trading customers, explained how he helps customers tune more efficiently with these SigOpt features in real-world scenarios.

Tuning for Systematic Trading: Talk 3: Training, Tuning, and Metric Strategy from SigOpt

]]> 175 0 https://cdn.slidesharecdn.com/ss_thumbnails/talkingseries-presentation1-200519230409-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Tuning for Systematic Trading: Talk 2: Deep Learning /slideshow/tuning-for-systematic-trading-talk-2-deep-learning/232433431 quantseries-4-200422175240
This talk explains how to train deep learning and other expensive models with parallelism and multitask optimization to reduce wall clock time. Tobias Andreassen, who supports a number of our systematic trading customers, presented the intuition behind Bayesian optimization for model optimization with a single or multiple (often competing) metrics. Many times it makes sense to analyze a second metric to avoid myopic training runs that overfit on your data, or otherwise don’t represent or impede performance in real-world scenarios.]]>
This talk explains how to train deep learning and other expensive models with parallelism and multitask optimization to reduce wall clock time. Tobias Andreassen, who supports a number of our systematic trading customers, presented the intuition behind Bayesian optimization for model optimization with a single or multiple (often competing) metrics. Many times it makes sense to analyze a second metric to avoid myopic training runs that overfit on your data, or otherwise don’t represent or impede performance in real-world scenarios.]]> Wed, 22 Apr 2020 17:52:40 GMT /slideshow/tuning-for-systematic-trading-talk-2-deep-learning/232433431 SigOpt@slideshare.net(SigOpt) Tuning for Systematic Trading: Talk 2: Deep Learning SigOpt This talk explains how to train deep learning and other expensive models with parallelism and multitask optimization to reduce wall clock time. Tobias Andreassen, who supports a number of our systematic trading customers, presented the intuition behind Bayesian optimization for model optimization with a single or multiple (often competing) metrics. Many times it makes sense to analyze a second metric to avoid myopic training runs that overfit on your data, or otherwise don’t represent or impede performance in real-world scenarios. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/quantseries-4-200422175240-thumbnail.jpg?width=120&height=120&fit=bounds" /> This talk explains how to train deep learning and other expensive models with parallelism and multitask optimization to reduce wall clock time. Tobias Andreassen, who supports a number of our systematic trading customers, presented the intuition behind Bayesian optimization for model optimization with a single or multiple (often competing) metrics. Many times it makes sense to analyze a second metric to avoid myopic training runs that overfit on your data, or otherwise don’t represent or impede performance in real-world scenarios.

Tuning for Systematic Trading: Talk 2: Deep Learning from SigOpt

]]> 119 1 https://cdn.slidesharecdn.com/ss_thumbnails/quantseries-4-200422175240-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Tuning for Systematic Trading: Talk 1 /slideshow/tuning-for-systematic-trading-talk-1/230852125 tuningforsystematictradingtalk1-200325162437
This talk discusses the intuition behind Bayesian optimization with and without multiple metrics. Tobias Andreassen, who supports a number of our systematic trading customers, presented the intuition behind Bayesian optimization for model optimization with a single or multiple (often competing) metrics. Many times it makes sense to analyze a second metric to avoid myopic training runs that overfit on your data, or otherwise don’t represent or impede performance in real-world scenarios.]]>
This talk discusses the intuition behind Bayesian optimization with and without multiple metrics. Tobias Andreassen, who supports a number of our systematic trading customers, presented the intuition behind Bayesian optimization for model optimization with a single or multiple (often competing) metrics. Many times it makes sense to analyze a second metric to avoid myopic training runs that overfit on your data, or otherwise don’t represent or impede performance in real-world scenarios.]]> Wed, 25 Mar 2020 16:24:37 GMT /slideshow/tuning-for-systematic-trading-talk-1/230852125 SigOpt@slideshare.net(SigOpt) Tuning for Systematic Trading: Talk 1 SigOpt This talk discusses the intuition behind Bayesian optimization with and without multiple metrics. Tobias Andreassen, who supports a number of our systematic trading customers, presented the intuition behind Bayesian optimization for model optimization with a single or multiple (often competing) metrics. Many times it makes sense to analyze a second metric to avoid myopic training runs that overfit on your data, or otherwise don’t represent or impede performance in real-world scenarios. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/tuningforsystematictradingtalk1-200325162437-thumbnail.jpg?width=120&height=120&fit=bounds" /> This talk discusses the intuition behind Bayesian optimization with and without multiple metrics. Tobias Andreassen, who supports a number of our systematic trading customers, presented the intuition behind Bayesian optimization for model optimization with a single or multiple (often competing) metrics. Many times it makes sense to analyze a second metric to avoid myopic training runs that overfit on your data, or otherwise don’t represent or impede performance in real-world scenarios.

Tuning for Systematic Trading: Talk 1 from SigOpt

]]> 191 0 https://cdn.slidesharecdn.com/ss_thumbnails/tuningforsystematictradingtalk1-200325162437-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Tuning Data Augmentation to Boost Model Performance /SigOpt/tuning-data-augmentation-to-boost-model-performance meghanasimageaugmentationwebinar2019-191205223609
In this webinar, SigOpt ML Engineer Meghana Ravikumar presents on and builds an image classifier trained on the Stanford Cars dataset to evaluate two approaches to transfer learning—fine tuning and feature extraction—and the impact of Multitask optimization, a more efficient form of Bayesian optimization, on these techniques. Once we define the most performant transfer learning technique for Stanford Cars, we will use image augmentation to double the size of the dataset to boost the classifier’s performance. Instead of manually tuning the hyperparameters associated with image augmentation, we will use Multitask Optimization to learn these hyperparameters using the downstream image classifier’s performance as the guide. In conjunction with model performance, we will also explore the features of these augmented images and the downstream implications for our image classifier.]]>
In this webinar, SigOpt ML Engineer Meghana Ravikumar presents on and builds an image classifier trained on the Stanford Cars dataset to evaluate two approaches to transfer learning—fine tuning and feature extraction—and the impact of Multitask optimization, a more efficient form of Bayesian optimization, on these techniques. Once we define the most performant transfer learning technique for Stanford Cars, we will use image augmentation to double the size of the dataset to boost the classifier’s performance. Instead of manually tuning the hyperparameters associated with image augmentation, we will use Multitask Optimization to learn these hyperparameters using the downstream image classifier’s performance as the guide. In conjunction with model performance, we will also explore the features of these augmented images and the downstream implications for our image classifier.]]> Thu, 05 Dec 2019 22:36:09 GMT /SigOpt/tuning-data-augmentation-to-boost-model-performance SigOpt@slideshare.net(SigOpt) Tuning Data Augmentation to Boost Model Performance SigOpt In this webinar, SigOpt ML Engineer Meghana Ravikumar presents on and builds an image classifier trained on the Stanford Cars dataset to evaluate two approaches to transfer learning—fine tuning and feature extraction—and the impact of Multitask optimization, a more efficient form of Bayesian optimization, on these techniques. Once we define the most performant transfer learning technique for Stanford Cars, we will use image augmentation to double the size of the dataset to boost the classifier’s performance. Instead of manually tuning the hyperparameters associated with image augmentation, we will use Multitask Optimization to learn these hyperparameters using the downstream image classifier’s performance as the guide. In conjunction with model performance, we will also explore the features of these augmented images and the downstream implications for our image classifier. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/meghanasimageaugmentationwebinar2019-191205223609-thumbnail.jpg?width=120&height=120&fit=bounds" /> In this webinar, SigOpt ML Engineer Meghana Ravikumar presents on and builds an image classifier trained on the Stanford Cars dataset to evaluate two approaches to transfer learning—fine tuning and feature extraction—and the impact of Multitask optimization, a more efficient form of Bayesian optimization, on these techniques. Once we define the most performant transfer learning technique for Stanford Cars, we will use image augmentation to double the size of the dataset to boost the classifier’s performance. Instead of manually tuning the hyperparameters associated with image augmentation, we will use Multitask Optimization to learn these hyperparameters using the downstream image classifier’s performance as the guide. In conjunction with model performance, we will also explore the features of these augmented images and the downstream implications for our image classifier.

Tuning Data Augmentation to Boost Model Performance from SigOpt

]]> 162 1 https://cdn.slidesharecdn.com/ss_thumbnails/meghanasimageaugmentationwebinar2019-191205223609-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Advanced Optimization for the Enterprise Webinar /slideshow/advanced-optimization-for-the-enterprise-webinar/190838764 advancedoptimizationwebinarnov5-191106011149
Building on the TWIML eBook, TWIMLcon event and TWIML podcast series that explore Machine Learning Platforms in great detail, this webinar examines the machine learning platforms that power enterprise leaders in AI. SigOpt CEO Scott Clark will provide an overview of critical technical capabilities that our customers have prioritized in their ML platforms. Review these slides to learn about: - Critical capabilities for data, experiment and model management - Tradeoffs between building and buying these capabilities - Lessons from the implementation of these platforms by AI leaders Why focus on these platforms and the capabilities that power them? Nearly every company is investing in machine learning that differentiates products or generates revenue. These so-called "differentiated models" represent the biggest opportunity for AI to transform the business. Most of these teams find success hiring expert data scientists and machine learning engineers who can build these models. But most of these teams also struggle to create a more sustainable, scalable and reproducible process for model development, and have begun building ML platforms to tackle this challenge.]]>
Building on the TWIML eBook, TWIMLcon event and TWIML podcast series that explore Machine Learning Platforms in great detail, this webinar examines the machine learning platforms that power enterprise leaders in AI. SigOpt CEO Scott Clark will provide an overview of critical technical capabilities that our customers have prioritized in their ML platforms. Review these slides to learn about: - Critical capabilities for data, experiment and model management - Tradeoffs between building and buying these capabilities - Lessons from the implementation of these platforms by AI leaders Why focus on these platforms and the capabilities that power them? Nearly every company is investing in machine learning that differentiates products or generates revenue. These so-called "differentiated models" represent the biggest opportunity for AI to transform the business. Most of these teams find success hiring expert data scientists and machine learning engineers who can build these models. But most of these teams also struggle to create a more sustainable, scalable and reproducible process for model development, and have begun building ML platforms to tackle this challenge.]]> Wed, 06 Nov 2019 01:11:49 GMT /slideshow/advanced-optimization-for-the-enterprise-webinar/190838764 SigOpt@slideshare.net(SigOpt) Advanced Optimization for the Enterprise Webinar SigOpt Building on the TWIML eBook, TWIMLcon event and TWIML podcast series that explore Machine Learning Platforms in great detail, this webinar examines the machine learning platforms that power enterprise leaders in AI. SigOpt CEO Scott Clark will provide an overview of critical technical capabilities that our customers have prioritized in their ML platforms. Review these slides to learn about: - Critical capabilities for data, experiment and model management - Tradeoffs between building and buying these capabilities - Lessons from the implementation of these platforms by AI leaders Why focus on these platforms and the capabilities that power them? Nearly every company is investing in machine learning that differentiates products or generates revenue. These so-called "differentiated models" represent the biggest opportunity for AI to transform the business. Most of these teams find success hiring expert data scientists and machine learning engineers who can build these models. But most of these teams also struggle to create a more sustainable, scalable and reproducible process for model development, and have begun building ML platforms to tackle this challenge. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/advancedoptimizationwebinarnov5-191106011149-thumbnail.jpg?width=120&height=120&fit=bounds" /> Building on the TWIML eBook, TWIMLcon event and TWIML podcast series that explore Machine Learning Platforms in great detail, this webinar examines the machine learning platforms that power enterprise leaders in AI. SigOpt CEO Scott Clark will provide an overview of critical technical capabilities that our customers have prioritized in their ML platforms. Review these slides to learn about: - Critical capabilities for data, experiment and model management - Tradeoffs between building and buying these capabilities - Lessons from the implementation of these platforms by AI leaders Why focus on these platforms and the capabilities that power them? Nearly every company is investing in machine learning that differentiates products or generates revenue. These so-called "differentiated models" represent the biggest opportunity for AI to transform the business. Most of these teams find success hiring expert data scientists and machine learning engineers who can build these models. But most of these teams also struggle to create a more sustainable, scalable and reproducible process for model development, and have begun building ML platforms to tackle this challenge.

Advanced Optimization for the Enterprise Webinar from SigOpt

]]> 186 4 https://cdn.slidesharecdn.com/ss_thumbnails/advancedoptimizationwebinarnov5-191106011149-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Modeling at Scale: SigOpt at TWIMLcon 2019 /slideshow/modeling-at-scale-sigopt-at-twimlcon-2019/180450163 twimlcon19sigoptpresentation-modelingatscale-oct2019printver-191009174753
SigOpt founder and CEO, Scott Clark, PhD, explains the tradeoffs you'll want to consider when designing your modeling platform and integrating hyperparameter optimization to enhance data scientist productivity.]]>
SigOpt founder and CEO, Scott Clark, PhD, explains the tradeoffs you'll want to consider when designing your modeling platform and integrating hyperparameter optimization to enhance data scientist productivity.]]> Wed, 09 Oct 2019 17:47:53 GMT /slideshow/modeling-at-scale-sigopt-at-twimlcon-2019/180450163 SigOpt@slideshare.net(SigOpt) Modeling at Scale: SigOpt at TWIMLcon 2019 SigOpt SigOpt founder and CEO, Scott Clark, PhD, explains the tradeoffs you'll want to consider when designing your modeling platform and integrating hyperparameter optimization to enhance data scientist productivity. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/twimlcon19sigoptpresentation-modelingatscale-oct2019printver-191009174753-thumbnail.jpg?width=120&height=120&fit=bounds" /> SigOpt founder and CEO, Scott Clark, PhD, explains the tradeoffs you'll want to consider when designing your modeling platform and integrating hyperparameter optimization to enhance data scientist productivity.

Modeling at Scale: SigOpt at TWIMLcon 2019 from SigOpt

]]> 817 3 https://cdn.slidesharecdn.com/ss_thumbnails/twimlcon19sigoptpresentation-modelingatscale-oct2019printver-191009174753-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Tuning 2.0: Advanced Optimization Techniques Webinar /slideshow/tuning-20-advanced-optimization-techniques-webinar/170702160 tuning2-190911003518
This webinar, hosted by SigOpt co-founder and CEO Scott Clark, explains how advanced features can help you achieve your modeling goals. These features include metric definition and multimetric optimization, conditional parameters, and multitask optimization for long training cycles.]]>
This webinar, hosted by SigOpt co-founder and CEO Scott Clark, explains how advanced features can help you achieve your modeling goals. These features include metric definition and multimetric optimization, conditional parameters, and multitask optimization for long training cycles.]]> Wed, 11 Sep 2019 00:35:18 GMT /slideshow/tuning-20-advanced-optimization-techniques-webinar/170702160 SigOpt@slideshare.net(SigOpt) Tuning 2.0: Advanced Optimization Techniques Webinar SigOpt This webinar, hosted by SigOpt co-founder and CEO Scott Clark, explains how advanced features can help you achieve your modeling goals. These features include metric definition and multimetric optimization, conditional parameters, and multitask optimization for long training cycles. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/tuning2-190911003518-thumbnail.jpg?width=120&height=120&fit=bounds" /> This webinar, hosted by SigOpt co-founder and CEO Scott Clark, explains how advanced features can help you achieve your modeling goals. These features include metric definition and multimetric optimization, conditional parameters, and multitask optimization for long training cycles.

Tuning 2.0: Advanced Optimization Techniques Webinar from SigOpt

]]> 714 4 https://cdn.slidesharecdn.com/ss_thumbnails/tuning2-190911003518-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 SigOpt at Ai4 Finance—Modeling at Scale /slideshow/sigopt-at-ai4-financemodeling-at-scale/166672847 ai4financesigoptpresentation-modelingatscale-august2019-190827003456
SigOpt helps your algorithmic traders and data scientists build better models faster. Learn how to integrate SigOpt into your modeling platform for quick ROI for your data science team.]]>
SigOpt helps your algorithmic traders and data scientists build better models faster. Learn how to integrate SigOpt into your modeling platform for quick ROI for your data science team.]]> Tue, 27 Aug 2019 00:34:56 GMT /slideshow/sigopt-at-ai4-financemodeling-at-scale/166672847 SigOpt@slideshare.net(SigOpt) SigOpt at Ai4 Finance—Modeling at Scale SigOpt SigOpt helps your algorithmic traders and data scientists build better models faster. Learn how to integrate SigOpt into your modeling platform for quick ROI for your data science team. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/ai4financesigoptpresentation-modelingatscale-august2019-190827003456-thumbnail.jpg?width=120&height=120&fit=bounds" /> SigOpt helps your algorithmic traders and data scientists build better models faster. Learn how to integrate SigOpt into your modeling platform for quick ROI for your data science team.

SigOpt at Ai4 Finance—Modeling at Scale from SigOpt

]]> 581 1 https://cdn.slidesharecdn.com/ss_thumbnails/ai4financesigoptpresentation-modelingatscale-august2019-190827003456-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimization /slideshow/interactive-tradeoffs-between-competing-offline-metrics-with-bayesian-optimization/163234594 michaelmccourt-sigopt-kddevaluationworkshop-interactivetradeoffsbetweencompetingofflinemetricswithba-190812171912
Many real world applications - machine learning models, simulators, etc. - have multiple competing metrics that define performance; these require practitioners to carefully consider potential tradeoffs. However, assessing and ranking this tradeoff is nontrivial, especially when the number of metrics is more than two. Often times, practitioners scalarize the metrics into a single objective, e.g., using a weighted sum. In this talk, we pose this problem as a constrained multi-objective optimization problem. By setting and updating the constraints, we can efficiently explore only the region of the Pareto efficient frontier of the model/system of most interest. We motivate this problem with the application of an experimental design setting, where we are trying to fabricate high performance glass substrate for solar cell panels. ]]>
Many real world applications - machine learning models, simulators, etc. - have multiple competing metrics that define performance; these require practitioners to carefully consider potential tradeoffs. However, assessing and ranking this tradeoff is nontrivial, especially when the number of metrics is more than two. Often times, practitioners scalarize the metrics into a single objective, e.g., using a weighted sum. In this talk, we pose this problem as a constrained multi-objective optimization problem. By setting and updating the constraints, we can efficiently explore only the region of the Pareto efficient frontier of the model/system of most interest. We motivate this problem with the application of an experimental design setting, where we are trying to fabricate high performance glass substrate for solar cell panels. ]]> Mon, 12 Aug 2019 17:19:12 GMT /slideshow/interactive-tradeoffs-between-competing-offline-metrics-with-bayesian-optimization/163234594 SigOpt@slideshare.net(SigOpt) Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimization SigOpt Many real world applications - machine learning models, simulators, etc. - have multiple competing metrics that define performance; these require practitioners to carefully consider potential tradeoffs. However, assessing and ranking this tradeoff is nontrivial, especially when the number of metrics is more than two. Often times, practitioners scalarize the metrics into a single objective, e.g., using a weighted sum. In this talk, we pose this problem as a constrained multi-objective optimization problem. By setting and updating the constraints, we can efficiently explore only the region of the Pareto efficient frontier of the model/system of most interest. We motivate this problem with the application of an experimental design setting, where we are trying to fabricate high performance glass substrate for solar cell panels. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/michaelmccourt-sigopt-kddevaluationworkshop-interactivetradeoffsbetweencompetingofflinemetricswithba-190812171912-thumbnail.jpg?width=120&height=120&fit=bounds" /> Many real world applications - machine learning models, simulators, etc. - have multiple competing metrics that define performance; these require practitioners to carefully consider potential tradeoffs. However, assessing and ranking this tradeoff is nontrivial, especially when the number of metrics is more than two. Often times, practitioners scalarize the metrics into a single objective, e.g., using a weighted sum. In this talk, we pose this problem as a constrained multi-objective optimization problem. By setting and updating the constraints, we can efficiently explore only the region of the Pareto efficient frontier of the model/system of most interest. We motivate this problem with the application of an experimental design setting, where we are trying to fabricate high performance glass substrate for solar cell panels.

Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimization from SigOpt

]]> 272 3 https://cdn.slidesharecdn.com/ss_thumbnails/michaelmccourt-sigopt-kddevaluationworkshop-interactivetradeoffsbetweencompetingofflinemetricswithba-190812171912-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Machine Learning Infrastructure /slideshow/machine-learning-infrastructure-149342518/149342518 pydataannarbor2019mlinfrastructure-190613014345
As data science workloads grow, so does their need for infrastructure. But, is it fair to ask data scientists to also become infrastructure experts? If not the data scientists, then, who is responsible for spinning up and managing data science infrastructure? This talk will address the context in which ML infrastructure is emerging, walk through two examples of ML infrastructure tools for launching hyperparameter optimization jobs, and end with some thoughts for building better tools in the future. Originally given as a talk at the PyData Ann Arbor meetup (https://www.meetup.com/PyData-Ann-Arbor/events/260380989/)]]>
As data science workloads grow, so does their need for infrastructure. But, is it fair to ask data scientists to also become infrastructure experts? If not the data scientists, then, who is responsible for spinning up and managing data science infrastructure? This talk will address the context in which ML infrastructure is emerging, walk through two examples of ML infrastructure tools for launching hyperparameter optimization jobs, and end with some thoughts for building better tools in the future. Originally given as a talk at the PyData Ann Arbor meetup (https://www.meetup.com/PyData-Ann-Arbor/events/260380989/)]]> Thu, 13 Jun 2019 01:43:45 GMT /slideshow/machine-learning-infrastructure-149342518/149342518 SigOpt@slideshare.net(SigOpt) Machine Learning Infrastructure SigOpt As data science workloads grow, so does their need for infrastructure. But, is it fair to ask data scientists to also become infrastructure experts? If not the data scientists, then, who is responsible for spinning up and managing data science infrastructure? This talk will address the context in which ML infrastructure is emerging, walk through two examples of ML infrastructure tools for launching hyperparameter optimization jobs, and end with some thoughts for building better tools in the future. Originally given as a talk at the PyData Ann Arbor meetup (https://www.meetup.com/PyData-Ann-Arbor/events/260380989/) <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/pydataannarbor2019mlinfrastructure-190613014345-thumbnail.jpg?width=120&height=120&fit=bounds" /> As data science workloads grow, so does their need for infrastructure. But, is it fair to ask data scientists to also become infrastructure experts? If not the data scientists, then, who is responsible for spinning up and managing data science infrastructure? This talk will address the context in which ML infrastructure is emerging, walk through two examples of ML infrastructure tools for launching hyperparameter optimization jobs, and end with some thoughts for building better tools in the future. Originally given as a talk at the PyData Ann Arbor meetup (https://www.meetup.com/PyData-Ann-Arbor/events/260380989/)

Machine Learning Infrastructure from SigOpt

]]> 637 2 https://cdn.slidesharecdn.com/ss_thumbnails/pydataannarbor2019mlinfrastructure-190613014345-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimization through customer engagements /slideshow/sigopt-at-uber-science-symposium-exploring-the-spectrum-of-blackbox-optimization-through-customer-engagements/146277084 ubersymposiumslides-190517162441
At the inaugural Uber science symposium, SigOpt research engineer Bolong (Harvey) Cheng shares insights on black-box optimization from his experience working with both leading academics and innovative enterprises. ]]>
At the inaugural Uber science symposium, SigOpt research engineer Bolong (Harvey) Cheng shares insights on black-box optimization from his experience working with both leading academics and innovative enterprises. ]]> Fri, 17 May 2019 16:24:41 GMT /slideshow/sigopt-at-uber-science-symposium-exploring-the-spectrum-of-blackbox-optimization-through-customer-engagements/146277084 SigOpt@slideshare.net(SigOpt) SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimization through customer engagements SigOpt At the inaugural Uber science symposium, SigOpt research engineer Bolong (Harvey) Cheng shares insights on black-box optimization from his experience working with both leading academics and innovative enterprises. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/ubersymposiumslides-190517162441-thumbnail.jpg?width=120&height=120&fit=bounds" /> At the inaugural Uber science symposium, SigOpt research engineer Bolong (Harvey) Cheng shares insights on black-box optimization from his experience working with both leading academics and innovative enterprises.

SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimization through customer engagements from SigOpt

]]> 281 6 https://cdn.slidesharecdn.com/ss_thumbnails/ubersymposiumslides-190517162441-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 SigOpt at O'Reilly - Best Practices for Scaling Modeling Platforms /slideshow/sigopt-at-oreilly-best-practices-for-scaling-modeling-platforms/145897625 oreillyainyc2019bestpracticesforscalingmodelingplatforms-190515223702
Companies are increasingly building modeling platforms to empower their researchers to efficiently scale the development and productionalization of their models. Scott Clark and Matt Greenwood share a case study from a leading algorithmic trading firm to illustrate best practices for building these types of platforms in any industry. Join in to learn how Two Sigma, a leading quantitative investment and technology firm, solved its model optimization problem.]]>
Companies are increasingly building modeling platforms to empower their researchers to efficiently scale the development and productionalization of their models. Scott Clark and Matt Greenwood share a case study from a leading algorithmic trading firm to illustrate best practices for building these types of platforms in any industry. Join in to learn how Two Sigma, a leading quantitative investment and technology firm, solved its model optimization problem.]]> Wed, 15 May 2019 22:37:02 GMT /slideshow/sigopt-at-oreilly-best-practices-for-scaling-modeling-platforms/145897625 SigOpt@slideshare.net(SigOpt) SigOpt at O'Reilly - Best Practices for Scaling Modeling Platforms SigOpt Companies are increasingly building modeling platforms to empower their researchers to efficiently scale the development and productionalization of their models. Scott Clark and Matt Greenwood share a case study from a leading algorithmic trading firm to illustrate best practices for building these types of platforms in any industry. Join in to learn how Two Sigma, a leading quantitative investment and technology firm, solved its model optimization problem. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/oreillyainyc2019bestpracticesforscalingmodelingplatforms-190515223702-thumbnail.jpg?width=120&height=120&fit=bounds" /> Companies are increasingly building modeling platforms to empower their researchers to efficiently scale the development and productionalization of their models. Scott Clark and Matt Greenwood share a case study from a leading algorithmic trading firm to illustrate best practices for building these types of platforms in any industry. Join in to learn how Two Sigma, a leading quantitative investment and technology firm, solved its model optimization problem.

SigOpt at O'Reilly - Best Practices for Scaling Modeling Platforms from SigOpt

]]> 227 1 https://cdn.slidesharecdn.com/ss_thumbnails/oreillyainyc2019bestpracticesforscalingmodelingplatforms-190515223702-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 SigOpt at GTC - Tuning the Untunable /SigOpt/sigopt-at-gtc-tuning-the-untunable sigoptatgtcsj2019-tuningtheuntunable-190515220150
Training and tuning models with lengthy training cycles like those in deep learning can be extremely expensive and may sometimes involve techniques that degrade performance. We'll explore recent research on optimization strategies to efficiently tune these types of deep learning models. We will provide benchmarks and comparisons to other popular methods for optimizing the models, and we'll recommend valuable areas for further applied research.]]>
Training and tuning models with lengthy training cycles like those in deep learning can be extremely expensive and may sometimes involve techniques that degrade performance. We'll explore recent research on optimization strategies to efficiently tune these types of deep learning models. We will provide benchmarks and comparisons to other popular methods for optimizing the models, and we'll recommend valuable areas for further applied research.]]> Wed, 15 May 2019 22:01:49 GMT /SigOpt/sigopt-at-gtc-tuning-the-untunable SigOpt@slideshare.net(SigOpt) SigOpt at GTC - Tuning the Untunable SigOpt Training and tuning models with lengthy training cycles like those in deep learning can be extremely expensive and may sometimes involve techniques that degrade performance. We'll explore recent research on optimization strategies to efficiently tune these types of deep learning models. We will provide benchmarks and comparisons to other popular methods for optimizing the models, and we'll recommend valuable areas for further applied research. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/sigoptatgtcsj2019-tuningtheuntunable-190515220150-thumbnail.jpg?width=120&height=120&fit=bounds" /> Training and tuning models with lengthy training cycles like those in deep learning can be extremely expensive and may sometimes involve techniques that degrade performance. We'll explore recent research on optimization strategies to efficiently tune these types of deep learning models. We will provide benchmarks and comparisons to other popular methods for optimizing the models, and we'll recommend valuable areas for further applied research.

SigOpt at GTC - Tuning the Untunable from SigOpt

]]> 228 4 https://cdn.slidesharecdn.com/ss_thumbnails/sigoptatgtcsj2019-tuningtheuntunable-190515220150-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 SigOpt at GTC - Reducing operational barriers to optimization /slideshow/sigopt-at-gtc-reducing-operational-barriers-to-optimization/145894182 sigoptatgtcsj2019-reducingoperationalbarrierstooptimization-190515215754
Advanced hardware like NVIDIA technology lowers technical barriers to model size and scope, but issues remain in areas like model performance and training infrastructure management. We'll discuss operational challenges to training models at scale with a particular focus on how training management and hyperparameter tuning can inform each other to accomplish specific goals. We'll also explore techniques like parallelism and scheduling, discuss their impact on model optimization, and compare various techniques. We'll also evaluate results of this approach. In particular, we'll focus on how new tools that automate training orchestration accelerate model development and increase the volume and quality of models in production.]]>
Advanced hardware like NVIDIA technology lowers technical barriers to model size and scope, but issues remain in areas like model performance and training infrastructure management. We'll discuss operational challenges to training models at scale with a particular focus on how training management and hyperparameter tuning can inform each other to accomplish specific goals. We'll also explore techniques like parallelism and scheduling, discuss their impact on model optimization, and compare various techniques. We'll also evaluate results of this approach. In particular, we'll focus on how new tools that automate training orchestration accelerate model development and increase the volume and quality of models in production.]]> Wed, 15 May 2019 21:57:54 GMT /slideshow/sigopt-at-gtc-reducing-operational-barriers-to-optimization/145894182 SigOpt@slideshare.net(SigOpt) SigOpt at GTC - Reducing operational barriers to optimization SigOpt Advanced hardware like NVIDIA technology lowers technical barriers to model size and scope, but issues remain in areas like model performance and training infrastructure management. We'll discuss operational challenges to training models at scale with a particular focus on how training management and hyperparameter tuning can inform each other to accomplish specific goals. We'll also explore techniques like parallelism and scheduling, discuss their impact on model optimization, and compare various techniques. We'll also evaluate results of this approach. In particular, we'll focus on how new tools that automate training orchestration accelerate model development and increase the volume and quality of models in production. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/sigoptatgtcsj2019-reducingoperationalbarrierstooptimization-190515215754-thumbnail.jpg?width=120&height=120&fit=bounds" /> Advanced hardware like NVIDIA technology lowers technical barriers to model size and scope, but issues remain in areas like model performance and training infrastructure management. We'll discuss operational challenges to training models at scale with a particular focus on how training management and hyperparameter tuning can inform each other to accomplish specific goals. We'll also explore techniques like parallelism and scheduling, discuss their impact on model optimization, and compare various techniques. We'll also evaluate results of this approach. In particular, we'll focus on how new tools that automate training orchestration accelerate model development and increase the volume and quality of models in production.

SigOpt at GTC - Reducing operational barriers to optimization from SigOpt

]]> 304 1 https://cdn.slidesharecdn.com/ss_thumbnails/sigoptatgtcsj2019-reducingoperationalbarrierstooptimization-190515215754-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Lessons for an enterprise approach to modeling at scale /slideshow/lessons-for-an-enterprise-approach-to-modeling-at-scale/145888510 lessonsforanenterpriseapproachtomodelingatscale1-190515210204
In this presentation, SigOpt draws lessons from working with leading modeling-driven enterprises to scale their development of high-performing models. ]]>
In this presentation, SigOpt draws lessons from working with leading modeling-driven enterprises to scale their development of high-performing models. ]]> Wed, 15 May 2019 21:02:04 GMT /slideshow/lessons-for-an-enterprise-approach-to-modeling-at-scale/145888510 SigOpt@slideshare.net(SigOpt) Lessons for an enterprise approach to modeling at scale SigOpt In this presentation, SigOpt draws lessons from working with leading modeling-driven enterprises to scale their development of high-performing models. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/lessonsforanenterpriseapproachtomodelingatscale1-190515210204-thumbnail.jpg?width=120&height=120&fit=bounds" /> In this presentation, SigOpt draws lessons from working with leading modeling-driven enterprises to scale their development of high-performing models.

Lessons for an enterprise approach to modeling at scale from SigOpt

]]> 168 1 https://cdn.slidesharecdn.com/ss_thumbnails/lessonsforanenterpriseapproachtomodelingatscale1-190515210204-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 https://cdn.slidesharecdn.com/profile-photo-SigOpt-48x48.jpg?cb=1641834940 SigOpt is the optimization platform that amplifies your research. SigOpt takes any research pipeline and tunes it, right in place. Our cloud-based ensemble of optimization algorithms is proven and seamless to deploy, and is used by globally recognized leaders within the insurance, credit card, algorithmic trading and consumer packaged goods industries. sigopt.com https://cdn.slidesharecdn.com/ss_thumbnails/multimetricbertf9webinar-200820220109-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/optimizing-bert-and-natural-language-models-with-sigopt-experiment-management/238084938 Optimizing BERT and Na... https://cdn.slidesharecdn.com/ss_thumbnails/f9webinar-sigopt-mlopssolutionfortheenterprise-200710232253-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/experiment-management-for-the-enterprise/236800355 Experiment Management ... https://cdn.slidesharecdn.com/ss_thumbnails/webinarmultimetricbert-200625234510-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/efficient-nlp-by-distilling-bert-and-multimetric-optimization/236223976 Efficient NLP by Disti...