AI News, Q&A with Sentient: on Artificial Intelligence, Open Source Machine Learning Model Management Framework, and the Studio.ML Platform

Q&A with Sentient: on Artificial Intelligence, Open Source Machine Learning Model Management Framework, and the Studio.ML Platform

Reposted from SourceForge.net The meteoric rise of artificial intelligence (AI), particularly machine learning (ML), is significantly disrupting the business landscape and spurring workforce change across different industries such as manufacturing, transportation, retailing, finance, healthcare, advertising, insurance, entertainment, and more.

As businesses enter into a new era ruled by data, organizations can now leverage machine learning algorithms to identify key trends and gain actionable insights by analyzing vast reams of data to make faster and smarter business decisions.

With virtually every industry transforming their business models and core processes to take advantage of machine learning tools, companies should invest in a unifying framework that can help them solve data problems and lessen overhead involved with running, scheduling, monitoring, and managing their machine learning experiments.

Navruzyan also discusses the benefits of using machine learning model management frameworks and shares how Studio.ML can support data scientists and DevOps to accelerate their scientific research and produce models that help analyze complex and large volumes of data in order to deliver faster, more accurate results.

First, digital marketing, through a product called Sentient Ascend, which provides a new approach to conversion rate optimization (CRO) using genetic algorithms to automate and accelerate the testing of thousands of web page combinations in the same time as a single A/B test.

We’ll be able to leverage AI systems to help get things to where they need to go faster and cheaper, we’ll be able to enable people to buy things they weren’t even aware existed or even knew they wanted, we’ll be able to help predict fatal diseases before they get past the point of no return.

The role of the ML model management frameworks is, on one hand, to partially automate the trial-and-error via so-called hyperparameter optimization or search (in essence, search for what training options work best for a given problem), and on the other hand, to keep track of what has been tried and help researchers reuse those results (both their own and the colleagues’ results) – either in a form of model artifacts that can be pushed into production, or, e.g.

A: We believe that the biggest impact in AI will be made by automated approaches to model design (so-called AI creating AI – approaches such as reinforcement learning of neural network architecture or neuroevolution) and by aggregation of knowledge from multiple domains and datasets (transfer learning) that should help in the situations when high-quality labelled data is not available or too expensive.

For example, when captioning images for visually impaired people, direct training data (images with proper descriptions) is relatively scarce, however, one has a virtually infinite supply of text or images with simple attributes from the internet that transfer learning should be able to tap into.

At Sentient, our research team is heavily focused on both architecture search using neuroevolution (in fact, there have been several recent breakthroughs in that area whereby neuroevolution was able to find seemingly strange neural network architectures that nonetheless performed better than a hand-crafted state of the art), and on transfer learning.

While (fortunately) these are steadily getting cheaper by virtue of Moore’s law, managing an increasing amount of compute resources can generate an increasing amount hassle, so the role of machine learning management frameworks like Studio.ML will increase.

QA with Sentient: on Artificial Intelligence, Open Source Machine Learning Model Management Framework, and the Studio.ML Platform

The meteoric rise of artificial intelligence (AI), particularly machine learning (ML), is significantly disrupting the business landscape and spurring workforce change across different industries such as manufacturing, transportation, retailing, finance, healthcare, advertising, insurance, entertainment, and more.

As businesses enter into a new era ruled by data, organizations can now leverage machine learning algorithms to identify key trends and gain actionable insights by analyzing vast reams of data to make faster and smarter business decisions.

With virtually every industry transforming their business models and core processes to take advantage of machine learning tools, companies should invest in a unifying framework that can help them solve data problems and lessen overhead involved with running, scheduling, monitoring, and managing their machine learning experiments.

Navruzyan also discusses the benefits of using machine learning model management frameworks and shares how Studio.ML can support data scientists and DevOps to accelerate their scientific research and produce models that help analyze complex and large volumes of data in order to deliver faster, more accurate results.

First, digital marketing, through a product called Sentient Ascend, which provides a new approach to conversion rate optimization (CRO) using genetic algorithms to automate and accelerate the testing of thousands of web page combinations in the same time as a single A/B test.

Second, in retail with a product called Sentient Aware, which provides AI-powered product recommendations using deep learning and online learning to curate products for each customer based on their individual, in-the-moment visual preferences.

We’ll be able to leverage AI systems to help get things to where they need to go faster and cheaper, we’ll be able to enable people to buy things they weren’t even aware existed or even knew they wanted, we’ll be able to help predict fatal diseases before they get past the point of no return.

The role of the ML model management frameworks is, on one hand, to partially automate the trial-and-error via so-called hyperparameter optimization or search (in essence, search for what training options work best for a given problem), and on the other hand, to keep track of what has been tried and help researchers reuse those results (both their own and the colleagues’ results) –

approaches such as reinforcement learning of neural network architecture or neuroevolution) and by aggregation of knowledge from multiple domains and datasets (transfer learning) that should help in the situations when high-quality labelled data is not available or too expensive.

For example, when captioning images for visually impaired people, direct training data (images with proper descriptions) is relatively scarce, however, one has a virtually infinite supply of text or images with simple attributes from the internet that transfer learning should be able to tap into.

At Sentient, our research team is heavily focused on both architecture search using neuroevolution (in fact, there have been several recent breakthroughs in that area whereby neuroevolution was able to find seemingly strange neural network architectures that nonetheless performed better than a hand-crafted state of the art), and on transfer learning.

While (fortunately) these are steadily getting cheaper by virtue of Moore’s law, managing an increasing amount of compute resources can generate an increasing amount hassle, so the role of machine learning management frameworks like Studio.ML will increase.

Introducing Studio.ML: an Open Source Framework that Simplifies and Expedites Machine Learning Model Development

We’re big fans of Keras, TensorFlow, PyTorch etc., but along the way we’ve found some critical missing features when it comes to running a large set of machine learning experiments in a scalable and collaborative fashion.

Today, we’re excited to introduce Studio.ml, an open source project dedicated to helping machine learning (ML) professionals, academics, businesses and anyone else interested in ML model building, accelerate and simplify their experiments.

Studio.ml is an early-stage, ML model management framework written in Python and developed to minimize the overhead involved with scheduling, running, monitoring and managing artifacts of your machine learning experiments. Most

So far, using Studio.ml you can: This code is still in the early phases of development, but we encourage you to try it, report back problems, ask questions, provide feedback and contribute.

Running an experiment with Studio should be as simple as replacing `python` with `studio run` in your command line with a few flags for capturing your workspace or naming your experiments.

If everyone in your team uses Studio’s remote workers to launch and manage experiments, then you can view the experiments of other users through the web interface.

Studio.ML Machine Learning Meetup @ SentientDAI

Studio.ml is an early-stage, ML model management framework written in Python and developed to minimize the overhead involved with scheduling, running, monitoring and managing artifacts of your machine learning experiments.Most of the features are compatible with any Python ML framework, including Keras, TensorFlow, PyTorch, and scikit-learn (additional features available for Keras and TensorFlow).So far, using Studio.ml you can:Capture experiment information- Python environment, files, dependencies and logs- without modifying the experiment codeMonitor and organize experiments using a web dashboard that integrates with TensorBoardRun experiments locally, remotely, or in the cloud (Google Cloud or Amazon EC2)Manage artifactsPerform hyperparameter searchCreate customizable Python environments for remote executionAccess the model library to reuse models that have already been createdThis code is still in the early phases of development, but we encourage you to try it, report back problems, ask questions, provide feedback and contribute.

Zero-overhead scalable machine learning

By Peter Zhokhov, Senior Data Scientist We analyze the complexity overhead and a learning curve associated with the transition from quick-and-dirty machine learning experiments to large-scale production-grade models with the recently released Amazon SageMaker and open-source project Studio.ML.

With the number of machine learning experiments and models growing, data science teams in big and small companies realize the need for a unifying framework, one that lets data scientists build on top of their own models and experiments as well as leverage the efforts of other community teams and members in an efficient manner.

concept, however, faces multiple challenges (mainly related to the usage of large datasets, large amounts of computing resources and custom hardware).

In this blog series, we’ll build several models ranging from very simple toy examples to state-of-the-art deep neural networks presented at NIPS 2017.

The story revolves around a fairly simple exercise (chosen from the SageMaker getting started guide to ensure my lack of knowledge with the SageMaker is not affecting the results) –

Each step consists of assigning data samples (in our case, 28×28 grayscale MNIST images -> 784 dimensional vector samples) to clusters, and then moving the cluster centers to the averaged coordinates of data samples in each cluster.

The first cell deals with imports and downloading the MNIST data, while the second cell converts and uploads the data to the s3 bucket (the SageMaker training routine assumes data resides within S3).

Studio.ML gives you all of the above (serving, custom hardware in the cloud, hyperparameter optimization, model provenance) without leaving the comfort zone of the locally (or where ever your preference is)-running jupyter notebook or python command line.

K-Means with Studio.ML Let us install studioml package via pip install studioml Then, we use the same jupyter notebook as in the last exercise (K-means with sklearn), and add a single line to the imports section that imports cell magics from studio:

To start an experiment with studio, we simply add a cell magic %%studio_run to the notebook cell: Technically, studio.ml returns all the variables created in the cell, so I also erase train_set and test_set variables to prevent them being returned)

The link in the beginning of the experiment sends us to a central repo of experiments, to the experiment page that shows experiment progress, artifacts, and list of python packages necessary to reproduce the experiment.

The code in the cell runs in 6 minutes (slightly longer than plain sklearn due to, mainly, compressing, storing in the cloud, and returning the validation data;

The cool part is that once the training is complete, the rest of the notebook code is exactly the same as it used to be, including prediction and displaying cluster centers.

We then run the following command: studio serve 1513115524_jupyter_7afb38f0-9918-48b8-9921-0d29f44f421d –wrapper=kmeans_serve.py This command will serve the model locally, so in our notebook the following command generates the predictions:

In contrast, Studio.ML can seamlessly extend jupyter notebooks to provide the experiment provenance, training and / or inference on custom cloud compute (including spot instances) and serving.

The Fader Networks require a lot of computational resources to train, and as such, they are a tempting and yet difficult to chew fruit for data scientists outside large corporations such as Google or Facebook, which makes them a good demo for the full power of machine learning provenance frameworks such as the SageMaker or Studio.ML.

Getting started with Studio.ML

Studio.ML is a Python machine learning model management framework developed to help schedule, run, monitor and manage artifacts of your ML models.

Studio.ML Machine Learning Meetup @ SentientDAI

Studio.ml is an early-stage, ML model management framework written in Python and developed to minimize the overhead involved with scheduling, running, ...

Introduction to Studio.ML

Studio.ml is an early-stage, ML model management framework written in Python and developed to minimize the overhead involved with scheduling, running, ...

Introducing Seldon Deploy

We're excited to start lifting the lid on Seldon Deploy — a new enterprise product for data science teams that will be entering closed beta this summer. Seldon ...

Top 10 movies on Data Science & Machine Learning (2018)

If you are a good data analyst and movie lover, then this is one stop solution for your hunger of inspiration. Data enthusiast must see how the Artificial ...

UK TechDays Online is back!

This summer, we're setting up studio at the Microsoft Reactor in London and broadcasting through London Tech Week, bringing you a mix of deep technical ...

IBM Research Artificial Intelligence David Ferrucci

David Ferrucci, IBM Fellow and principal investigator on the IBM Watson Project, discussed what is on the horizon for the fields of artificial intelligence and ...

How ASU Cut Manual Email Sends in Half with Automation

Arizona State University does a high volume of email marketing. We average over 50 emails per day, many are series containing 5 or more emails. Currently we ...

An overview of Cloud IoT Core (Google I/O '18)

Come learn about Google Cloud's new IoT Platform -- Cloud IoT Core. This session will dive deep into how to connect and configure IoT devices with Google ...