Tag Archives: Open source

Introducing container-diff, a tool for quickly comparing container images

The Google Container Tools team originally built container-diff, a new project to help uncover differences between container images, to aid our own development with containers. We think it can be useful for anyone building containerized software, so we’re excited to release it as open source to the development community.

Containers and the Dockerfile format help make customization of an application’s runtime environment more approachable and easier to understand. While this is a great advantage of using containers in software development, a major drawback is that it can be hard to visualize what changes in a container image will result from a change in the respective Dockerfile. This can lead to bloated images and make tracking down issues difficult.

Imagine a scenario where a developer is working on an application, built on a runtime image maintained by a third-party. During development someone releases a new version of that base image with updated system packages. The developer rebuilds their application and picks up the latest version of the base image, and suddenly their application stops working; it depended on a previous version of one of the installed system packages, but which one? What version was it on before? With no currently existing tool to easily determine what changed between the two base image versions, this totally stalls development until the developer can track down the package version incompatibility.

Introducing container-diff

container-diff helps users investigate image changes by computing semantic diffs between images. What this means is that container-diff figures out on a low-level what data changed, and then combines this with an understanding of package manager information to output this information in a format that’s actually readable to users. The tool can find differences in system packages, language-level packages, and files in a container image.

Users can specify images in several formats - from local Docker daemon (using the prefix `daemon://` on the image path), a remote registry (using the prefix `remote://`), or a file in the .tar in the format exported by "docker save" command. You can also combine these formats to compute the diff between a local version of an image and a remote version. This can be useful when experimenting with new builds of an image that you might not be quite ready to push yet. container-diff supports image tarballs and the registry protocol natively, enabling it to run in environments without a Docker daemon.

Examples and Use Cases

Here is a basic Dockerfile that installs Python inside our Debian base image. Running container-diff on the base image and the new one with Python, users can see all the apt packages that were installed as dependencies of Python.


And below is a Dockerfile that inherits from our Python base runtime image, and then installs the mock and six packages inside of it. Running container-diff with the pip differ, users can see all the Python packages that have either been installed or changed as a result of this:


This can be especially useful when it’s unclear which packages might have been installed or changed incidentally as a result of dependency management of Python modules.

These are just a few examples. The tool currently has support for Python and Node.js packages installed via pip and npm, respectively, as well as comparison of image filesystems and Docker history. In the future, we’d like to see support added for additional runtime and language differs, including Java, Go, and Ruby. External contributions are welcome! For more information on contributing to container-diff, see this how-to guide.

Now that we’ve seen container-diff compare two images in action, it’s easy to imagine how the tool may be integrated into larger workflows to aid in development:
  • Changelog generation: Given container-diff’s capacity to facilitate investigation of filesystem and package modifications, it can do most of the heavy lifting in discerning changes for automatic changelog generation for new releases of an image.
  • Continuous integration: As part of a CI system, users can leverage container-diff to catch potentially breaking filesystem changes resulting from a Dockerfile change in their builds.
container-diff’s default output mode is “human-readable,” but also supports output to JSON, allowing for easy automated parsing and processing by users.

Single Image Analysis

In addition to comparing two images, container-diff has the ability to analyze a single image on its own. This can enable users to get a quick glance at information about an image, such as its system and language-level package installations and filesystem contents.

Let’s take a look at our Debian base image again. We can use the tool to easily view a list of all packages installed in the image, along with each one’s installed version and size:


We could use this to verify compatibility with an application we’re building, or maybe sort the packages by size in another one of our images and see which ones are taking up the most space.

For more information about this tool as well as a breakdown with examples, uses, and inner workings of the tool, please take a look at documentation on our GitHub page. Happy diffing!

Special thanks to Colette Torres and Abby Tisdale, our software engineering interns who helped build the tool from the ground up.

By Nick Kubala, Container Tools team


On-Device Conversational Modeling with TensorFlow Lite



Earlier this year, we launched Android Wear 2.0 which featured the first "on-device" machine learning technology for smart messaging. This enabled cloud-based technologies like Smart Reply, previously available in Gmail, Inbox and Allo, to be used directly within any application for the first time, including third-party messaging apps, without ever having to connect to the cloud. So you can respond to incoming chat messages on the go, directly from your smartwatch.

Today, we announce TensorFlow Lite, TensorFlow’s lightweight solution for mobile and embedded devices. This framework is optimized for low-latency inference of machine learning models, with a focus on small memory footprint and fast performance. As part of the library, we have also released an on-device conversational model and a demo app that provides an example of a natural language application powered by TensorFlow Lite, in order to make it easier for developers and researchers to build new machine intelligence features powered by on-device inference. This model generates reply suggestions to input conversational chat messages, with efficient inference that can be easily plugged in to your chat application to power on-device conversational intelligence.

The on-device conversational model we have released uses a new ML architecture for training compact neural networks (as well as other machine learning models) based on a joint optimization framework, originally presented in ProjectionNet: Learning Efficient On-Device Deep Networks Using Neural Projections. This architecture can run efficiently on mobile devices with limited computing power and memory, by using efficient “projection” operations that transform any input to a compact bit vector representation — similar inputs are projected to nearby vectors that are dense or sparse depending on type of projection. For example, the messages “hey, how's it going?” and “How's it going buddy?”, might be projected to the same vector representation.

Using this idea, the conversational model combines these efficient operations at low computation and memory footprint. We trained this on-device model end-to-end using an ML framework that jointly trains two types of models — a compact projection model (as described above) combined with a trainer model. The two models are trained in a joint fashion, where the projection model learns from the trainer model — the trainer is characteristic of an expert and modeled using larger and more complex ML architectures, whereas the projection model resembles a student that learns from the expert. During training, we can also stack other techniques such as quantization or distillation to achieve further compression or selectively optimize certain portions of the objective function. Once trained, the smaller projection model is able to be used directly for inference on device.
For inference, the trained projection model is compiled into a set of TensorFlow Lite operations that have been optimized for fast execution on mobile platforms and executed directly on device. The TensorFlow Lite inference graph for the on-device conversational model is shown here.
TensorFlow Lite execution for the On-Device Conversational Model.
The open-source conversational model released today (along with code) was trained end-to-end using the joint ML architecture described above. Today’s release also includes a demo app, so you can easily download and try out one-touch smart replies on your mobile device. The architecture enables easy configuration for model size and prediction quality based on application needs. You can find a list of sample messages where this model does well here. The system can also fall back to suggesting replies from a fixed set that was learned and compiled from popular response intents observed in chat conversations. The underlying model is different from the ones Google uses for Smart Reply responses in its apps1.

Beyond Conversational Models
Interestingly, the ML architecture described above permits flexible choices for the underlying model. We also designed the architecture to be compatible with different machine learning approaches — for example, when used with TensorFlow deep learning, we learn a lightweight neural network (ProjectionNet) for the underlying model, whereas a different architecture (ProjectionGraph) represents the model using a graph framework instead of a neural network.

The joint framework can also be used to train lightweight on-device models for other tasks using different ML modeling architectures. As an example, we derived a ProjectionNet architecture that uses a complex feed-forward or recurrent architecture (like LSTM) for the trainer model coupled with a simple projection architecture comprised of dynamic projection operations and a few, narrow fully-connected layers. The whole architecture is trained end-to-end using backpropagation in TensorFlow and once trained, the compact ProjectionNet is directly used for inference. Using this method, we have successfully trained tiny ProjectionNet models that achieve significant reduction in model sizes (up to several orders of magnitude reduction) and high performance with respect to accuracy on multiple visual and language classification tasks (a few examples here). Similarly, we trained other lightweight models using our graph learning framework, even in semi-supervised settings.
ML architecture for training on-device models: ProjectionNet trained using deep learning (left), and ProjectionGraph trained using graph learning (right).
We will continue to improve and release updated TensorFlow Lite models in open-source. We think that the released model (as well as future models) learned using these ML architectures may be reused for many natural language and computer vision applications or plugged into existing apps for enabling machine intelligence. We hope that the machine learning and natural language processing communities will be able to build on these to address new problems and use-cases we have not yet conceived.

Acknowledgments
Yicheng Fan and Gaurav Nemade contributed immensely to this effort. Special thanks to Rajat Monga, Andre Hentz, Andrew Selle, Sarah Sirajuddin, and Anitha Vijayakumar from the TensorFlow team; Robin Dua, Patrick McGregor, Andrei Broder, Andrew Tomkins and the Google Expander team.



1 The released on-device model was trained to optimize for small size and low latency applications on mobile phones and wearables. Smart Reply predictions in Google apps, however are generated using larger, more complex models. In production systems, we also use multiple classifiers that are trained to detect inappropriate content and apply further filtering and tuning to optimize user experience and quality levels. We recommend that developers using the open-source TensorFlow Lite version also follow such practices for their end applications.

Tangent: Source-to-Source Debuggable Derivatives

Crossposted on the Google Research Blog

Tangent is a new, free, and open source Python library for automatic differentiation. In contrast to existing machine learning libraries, Tangent is a source-to-source system, consuming a Python function f and emitting a new Python function that computes the gradient of f. This allows much better user visibility into gradient computations, as well as easy user-level editing and debugging of gradients. Tangent comes with many more features for debugging and designing machine learning models.
This post gives an overview of the Tangent API. It covers how to use Tangent to generate gradient code in Python that is easy to interpret, debug and modify.

Neural networks (NNs) have led to great advances in machine learning models for images, video, audio, and text. The fundamental abstraction that lets us train NNs to perform well at these tasks is a 30-year-old idea called reverse-mode automatic differentiation (also known as backpropagation), which comprises two passes through the NN. First, we run a “forward pass” to calculate the output value of each node. Then we run a “backward pass” to calculate a series of derivatives to determine how to update the weights to increase the model’s accuracy.

Training NNs, and doing research on novel architectures, requires us to compute these derivatives correctly, efficiently, and easily. We also need to be able to debug these derivatives when our model isn’t training well, or when we’re trying to build something new that we do not yet understand. Automatic differentiation, or just “autodiff,” is a technique to calculate the derivatives of computer programs that denote some mathematical function, and nearly every machine learning library implements it.

Existing libraries implement automatic differentiation by tracing a program’s execution (at runtime, like TF Eager, PyTorch and Autograd) or by building a dynamic data-flow graph and then differentiating the graph (ahead-of-time, like TensorFlow). In contrast, Tangent performs ahead-of-time autodiff on the Python source code itself, and produces Python source code as its output.
As a result, you can finally read your automatic derivative code just like the rest of your program. Tangent is useful to researchers and students who not only want to write their models in Python, but also read and debug automatically-generated derivative code without sacrificing speed and flexibility.

You can easily inspect and debug your models written in Tangent, without special tools or indirection. Tangent works on a large and growing subset of Python, provides extra autodiff features other Python ML libraries don’t have, is high-performance, and is compatible with TensorFlow and NumPy.

Automatic differentiation of Python code

How do we automatically generate derivatives of plain Python code? Math functions like tf.exp or tf.log have derivatives, which we can compose to build the backward pass. Similarly, pieces of syntax, such as  subroutines, conditionals, and loops, also have backward-pass versions. Tangent contains recipes for generating derivative code for each piece of Python syntax, along with many NumPy and TensorFlow function calls.

Tangent has a one-function API:
import tangent
df = tangent.grad(f)
Here’s an animated graphic of what happens when we call tangent.grad on a Python function:
If you want to print out your derivatives, you can run
import tangent
df = tangent.grad(f, verbose=1)
Under the hood, tangent.grad first grabs the source code of the Python function you pass it. Tangent has a large library of recipes for the derivatives of Python syntax, as well as TensorFlow Eager functions. The function tangent.grad then walks your code in reverse order, looks up the matching backward-pass recipe, and adds it to the end of the derivative function. This reverse-order processing gives the technique its name: reverse-mode automatic differentiation.

The function df above only works for scalar (non-array) inputs. Tangent also supports
Although we started with TensorFlow Eager support, Tangent isn’t tied to one numeric library or another—we would gladly welcome pull requests adding PyTorch or MXNet derivative recipes.

Next Steps

Tangent is open source now at github.com/google/tangent. Go check it out for download and installation instructions. Tangent is still an experiment, so expect some bugs. If you report them to us on GitHub, we will do our best to fix them quickly.

We are working to add support in Tangent for more aspects of the Python language (e.g., closures, inline function definitions, classes, more NumPy and TensorFlow functions). We also hope to add more advanced automatic differentiation and compiler functionality in the future, such as automatic trade-off between memory and compute (Griewank and Walther 2000; Gruslys et al., 2016), more aggressive optimizations, and lambda lifting.

We intend to develop Tangent together as a community. We welcome pull requests with fixes and features. Happy deriving!

By Alex Wiltschko, Research Scientist, Google Brain Team

Acknowledgments

Bart van Merriënboer contributed immensely to all aspects of Tangent during his internship, and Dan Moldovan led TF Eager integration, infrastructure and benchmarking. Also, thanks to the Google Brain team for their support of this post and special thanks to Sanders Kleinfeld and Aleks Haecky for their valuable contribution for the technical aspects of the post.

Latest Innovations in TensorFlow Serving



Since initially open-sourcing TensorFlow Serving in February 2016, we’ve made some major enhancements. Let’s take a look back at where we started, review our progress, and share where we are headed next.

Before TensorFlow Serving, users of TensorFlow inside Google had to create their own serving system from scratch. Although serving might appear easy at first, one-off serving solutions quickly grow in complexity. Machine Learning (ML) serving systems need to support model versioning (for model updates with a rollback option) and multiple models (for experimentation via A/B testing), while ensuring that concurrent models achieve high throughput on hardware accelerators (GPUs and TPUs) with low latency. So we set out to create a single, general TensorFlow Serving software stack.

We decided to make it open-sourceable from the get-go, and development started in September 2015. Within a few months, we created the initial end-to-end working system and our open-source release in February 2016.

Over the past year and half, with the help of our users and partners inside and outside our company, TensorFlow Serving has advanced performance, best practices, and standards:
  • Out-of-the-box optimized serving and customizability: We now offer a pre-built canonical serving binary, optimized for modern CPUs with AVX, so developers don't need to assemble their own binary from our libraries unless they have exotic needs. At the same time, we added a registry-based framework, allowing our libraries to be used for custom (or even non-TensorFlow) serving scenarios.
  • Multi-model serving: Going from one model to multiple concurrently-served models presents several performance obstacles. We serve multiple models smoothly by (1) loading in isolated thread pools to avoid incurring latency spikes on other models taking traffic; (2) accelerating initial loading of all models in parallel upon server start-up; (3) multi-model batch interleaving to multiplex hardware accelerators (GPUs/TPUs).
  • Standardized model format: We added SavedModel to TensorFlow 1.0, giving the community a single standard model format that works across training and serving.
  • Easy-to-use inference APIs: We released easy-to-use APIs for common inference tasks (classification, regression) that we know work for a wide swathe of our applications. To support more advanced use-cases we support a lower-level tensor-based API (predict) and a new multi-inference API that enables multi-task modeling.
All of our work has been informed by close collaborations with: (a) Google’s ML SRE team, which helps ensure we are robust and meet internal SLAs; (b) other Google machine learning infrastructure teams including ads serving and TFX; (c) application teams such as Google Play; (d) our partners at the UC Berkeley RISE Lab, who explore complementary research problems with the Clipper serving system; (e) our open-source user base and contributors.

TensorFlow Serving is currently handling tens of millions of inferences per second for 1100+ of our own projects including Google’s Cloud ML Prediction. Our core serving code is available to all via our open-source releases.

Looking forward, our work is far from done and we are exploring several avenues of innovation. Today we are excited to share early progress in two experimental areas:
  • Granular batching: A key technique we employ to achieve high throughput on specialized hardware (GPUs and TPUs) is "batching": processing multiple examples jointly for efficiency. We are developing technology and best practices to improve batching to: (a) enable batching to target just the GPU/TPU portion of the computation, for maximum efficiency; (b) enable batching within recursive neural networks, used to process sequence data e.g. text and event sequences. We are experimenting with batching arbitrary sub-graphs using the Batch/Unbatch op pair.
  • Distributed model serving: We are looking at model sharding techniques as a means of handling models that are too large to fit on one server node or sharing sub-models in a memory-efficient way. We recently launched a 1TB+ model in production with good results, and hope to open-source this capability soon.
Thanks again to all of our users and partners who have contributed feedback, code and ideas. Join the project at: github.com/tensorflow/serving.

Welcoming 25 mentor organizations for Google Code-in 2017

We’re thrilled to introduce 25 open source organizations that are participating in Google Code-in 2017. The contest, now in its eighth year, offers 13-17 year old pre-university students an opportunity to learn and practice their skills while contributing to open source projects.

Google Code-in officially starts for students on November 28. Students are encouraged to learn about the participating organizations ahead of time and can get started by clicking on the links below:

  • Apertium: rule-based machine translation platform
  • BRL-CAD: computer graphics, 2D and 3D geometry modeling and computer-aided design (CAD)
  • Catrobat: visual programming for creating mobile games and animations
  • CCExtractor: open source tools for subtitle generation
  • CloudCV: building platforms for reproducible AI research
  • coala: a unified interface for linting and fixing code, regardless of the programming languages used
  • Drupal: content management platform
  • FOSSASIA: developing communities across all ages and borders to form a better future with Open Technologies and ICT
  • Haiku: operating system specifically targeting personal computing
  • JBoss Community: a community of projects around JBoss Middleware
  • LibreHealth: aiming to bring open source healthcare IT to all of humanity
  • Liquid Galaxy: an interactive, panoramic and immersive visualization tool
  • MetaBrainz: builds community maintained databases
  • Mifos Initiative: transforming the delivery of financial services to the poor and the unbanked
  • MovingBlocks: a Minecraft-inspired open source game
  • OpenMRS: open source medical records system for the world
  • OpenWISP: build and manage low cost networks such as public wifi
  • OSGeo: building open source geospatial tools
  • Sugar Labs: learning platform and activities for elementary education
  • SCoRe: research lab seeking sustainable solutions for problems faced by developing countries
  • Systers: community for women involved in technical aspects of computing
  • Ubuntu: an open source operating system
  • Wikimedia: non-profit foundation dedicated to bringing free content to the world, operating Wikipedia
  • XWiki: a web platform for developing collaborative applications using the wiki paradigm
  • Zulip: powerful, threaded open source group chat with apps for every major platform

These mentor organizations are hard at work creating thousands of tasks for students to work on, including code, documentation, user interface, quality assurance, outreach, research and training tasks. The contest officially starts for students on Tuesday, November 28th at 9:00am PST.

You can learn more about Google Code-in on the contest site where you’ll find Contest Rules, Frequently Asked Questions and Important Dates. There you’ll also find flyers and other helpful information including the Getting Started Guide. Our discussion mailing list is a great way to talk with other students, mentors and organization administrators about the contest.

By Josh Simmons, Google Open Source

GCP Podcast hits 100 episodes — here are the 10 most popular episodes



It was 2015, and we were having coffee in a Google cafeteria when we, Francesc and Mark, came up with the idea for the Google Cloud Platform Podcast. Shortly thereafter, the first episode “We Got a Podcast” was released on the internet. And now, after years of interviewing Google Cloud Platform (GCP) customers, product managers, engineers, developer experts, salespeople, support staff and more, we’ve released our 100th episode!

To celebrate this milestone, we invited as our guest the venerable Vint Cerf, who's one of the “Fathers of the Internet,” having co-designed the TCP/IP protocols and the architecture of the internet, and who is currently vice president and Chief Internet Evangelist at Google. We talked to him about the history of the internet, net neutrality, the next billion users, interplanetary networks and more. His interview is totally amazing and you should definitely listen to it.

It also got us thinking about all the other amazing interviews we’ve done over the years. If you aren’t familiar with the Google Cloud Platform Podcast — or if you just want to relive the glory — here are the top 10 most downloaded episodes since it was introduced:

10. #93 What's AI with Melanie Warrick 
Fellow Developer Advocate, Melanie talks to us all about the differences between ML and AI, as well as discussing applications and where she sees this technology going in the future.

9. #82 Prometheus with Julius Volz 
Julius, co-founder of open source monitoring platform Prometheus talks to us all about the history of the platform as well as its capabilities as a monitoring and alerting system.

8. #90 Office of the CTO with Greg DeMichillie 
Greg DeMichillie talks about how GCP interacts with enterprise customers through the Office of the CTO, and lessons that he learned from them.

7. #56 A Year in Review
A review of Mark and Francesc’s favorite moments of 2016 — a great historical snapshot of the technology of GCP in 2016.
6. #76 Kubernetes 1.6 with Daniel Smith
We’re joined by one of the engineers on Kubernetes, Daniel Smith, and we talk in depth about the new features of Kubernetes 1.6, and how developers can take advantage of them. (Mark really nerds out on this one).

5. #62 Cloud Spanner with Deepti Srivastava
Cloud Spanner has been a hugely popular product for GCP, and we talk to Product Manager Deepti Srivastava about why it’s so technologically amazing as well as a game changer for customers looking for a distributed, horizontally scalable database.

4. #71 Cloud Machine Learning Engine with Yufeng Guo 
Another Developer Advocate, Yufeng, comes onto the podcast to introduce us to machine learning, and discuss how Cloud Machine Learning Engine can be used to train, as well as host, TensorFlow models.

3. #75 Container Engine with Chen Goldberg 
We're delighted to have Chen, Google Engineering Director for Container Engine and Kubernetes, to talk with us about the history of Kubernetes and why it’s open source, as well the integrations between Container Engine and GCP that make it so special.

2. #91 The Future of Media with Machine Learning with Amit Pande 
This is a fantastic episode with Amit, Product Management Leader for Google Cloud, to provide practical examples of how machine learning could be applied to the media and entertainment industry. Many things were learned!

1. #88 Kubernetes 1.7 with Tim Hockin 
One of the engineers that started the Kubernetes project, Tim comes onto the podcast to tell us all about the new feature of Kubernetes 1.7. Also, Tim is a super nice guy.

Lastly, the Google Cloud Platform Podcast would like to say thank you to all you listeners that have downloaded the podcast, the interviewees that gave us their time and all the people behind the scenes that helped edit, promote, procure equipment and more. Without all of you, the 100th episode would never have been possible.

To listen to more GCP podcasts, visit gcppodcast.com, or search for it in your favorite podcast app. See you all next week!

Announcing OpenFermion: The Open Source Chemistry Package for Quantum Computers

Crossposted on the Google Research Blog

“The underlying physical laws necessary for the mathematical theory of a large part of physics and the whole of chemistry are thus completely known, and the difficulty is only that the exact application of these laws leads to equations much too complicated to be soluble.”
-Paul Dirac, Quantum Mechanics of Many-Electron Systems (1929)

In this passage, physicist Paul Dirac laments that while quantum mechanics accurately models all of chemistry, exactly simulating the associated equations appears intractably complicated. Not until 1982 would Richard Feynman suggest that instead of surrendering to the complexity of quantum mechanics, we might harness it as a computational resource. Hence, the original motivation for quantum computing: by operating a computer according to the laws of quantum mechanics, one could efficiently unravel exact simulations of nature. Such simulations could lead to breakthroughs in areas such as photovoltaics, batteries, new materials, pharmaceuticals and superconductivity. And while we do not yet have a quantum computer large enough to solve classically intractable problems in these areas, rapid progress is being made. Last year, Google published this paper detailing the first quantum computation of a molecule using a superconducting qubit quantum computer. Building on that work, the quantum computing group at IBM scaled the experiment to larger molecules, which made the cover of Nature last month.

Today, we announce the release of OpenFermion, the first open source platform for translating problems in chemistry and materials science into quantum circuits that can be executed on existing platforms. OpenFermion is a library for simulating the systems of interacting electrons (fermions) which give rise to the properties of matter. Prior to OpenFermion, quantum algorithm developers would need to learn a significant amount of chemistry and write a large amount of code hacking apart other codes to put together even the most basic quantum simulations. While the project began at Google, collaborators at ETH Zurich, Lawrence Berkeley National Labs, University of Michigan, Harvard University, Oxford University, Dartmouth College, Rigetti Computing and NASA all contributed to alpha releases. You can learn more details about this release in our paper, OpenFermion: The Electronic Structure Package for Quantum Computers.

One way to think of OpenFermion is as a tool for generating and compiling physics equations which describe chemical and material systems into representations which can be interpreted by a quantum computer1. The most effective quantum algorithms for these problems build upon and extend the power of classical quantum chemistry packages used and developed by research chemists across government, industry and academia. Accordingly, we are also releasing OpenFermion-Psi4 and OpenFermion-PySCF which are plugins for using OpenFermion in conjunction with the classical electronic structure packages Psi4 and PySCF.

The core OpenFermion library is designed in a quantum programming framework agnostic way to ensure compatibility with various platforms being developed by the community. This allows OpenFermion to support external packages which compile quantum assembly language specifications for diverse hardware platforms. We hope this decision will help establish OpenFermion as a community standard for putting quantum chemistry on quantum computers. To see how OpenFermion is used with diverse quantum programming frameworks, take a look at OpenFermion-ProjectQ and Forest-OpenFermion - plugins which link OpenFermion to the externally developed circuit simulation and compilation platforms known as ProjectQ and Forest.

The following workflow describes how a quantum chemist might use OpenFermion in order to simulate the energy surface of a molecule (for instance, by preparing the sort of quantum computation we described in our past blog post):
  1. The researcher initializes an OpenFermion calculation with specification of:
    • An input file specifying the coordinates of the nuclei in the molecule.
    • The basis set (e.g. cc-pVTZ) that should be used to discretize the molecule.
    • The charge and spin multiplicity (if known) of the system.
  1. The researcher uses the OpenFermion-Psi4 plugin or the OpenFermion-PySCF plugin to perform scalable classical computations which are used to optimally stage the quantum computation. For instance, one might perform a classical Hartree-Fock calculation to choose a good initial state for the quantum simulation.
  2. The researcher then specifies which electrons are most interesting to study on a quantum computer (known as an active space) and asks OpenFermion to map the equations for those electrons to a representation suitable for quantum bits, using one of the available procedures in OpenFermion, e.g. the Bravyi-Kitaev transformation.
  3. The researcher selects a quantum algorithm to solve for the properties of interest and uses a quantum compilation framework such as OpenFermion-ProjectQ to output the quantum circuit in assembly language which can be run on a quantum computer. If the researcher has access to a quantum computer, they then execute the experiment.
A few examples of what one might do with OpenFermion are demonstrated in ipython notebooks here, here and here. While quantum simulation is widely recognized as one of the most important applications of quantum computing in the near term, very few quantum computer scientists know quantum chemistry and even fewer chemists know quantum computing. Our hope is that OpenFermion will help to close the gap between these communities and bring the power of quantum computing to chemists and material scientists. If you’re interested, please checkout our GitHub repository - pull requests welcome! By Ryan Babbush and Jarrod McClean, Quantum Software Engineers, Quantum AI Team

1 If we may be allowed one sentence for the experts: the primary function of OpenFermion is to encode the electronic structure problem in second quantization defined by various basis sets and active spaces and then to transform those operators into spin Hamiltonians using various isomorphisms between qubit and fermion algebras.

Announcing OpenFermion: The Open Source Chemistry Package for Quantum Computers



“The underlying physical laws necessary for the mathematical theory of a large part of physics and the whole of chemistry are thus completely known, and the difficulty is only that the exact application of these laws leads to equations much too complicated to be soluble.”
-Paul Dirac, Quantum Mechanics of Many-Electron Systems (1929)

In this passage, physicist Paul Dirac laments that while quantum mechanics accurately models all of chemistry, exactly simulating the associated equations appears intractably complicated. Not until 1982 would Richard Feynman suggest that instead of surrendering to the complexity of quantum mechanics, we might harness it as a computational resource. Hence, the original motivation for quantum computing: by operating a computer according to the laws of quantum mechanics, one could efficiently unravel exact simulations of nature. Such simulations could lead to breakthroughs in areas such as photovoltaics, batteries, new materials, pharmaceuticals and superconductivity. And while we do not yet have a quantum computer large enough to solve classically intractable problems in these areas, rapid progress is being made. Last year, Google published this paper detailing the first quantum computation of a molecule using a superconducting qubit quantum computer. Building on that work, the quantum computing group at IBM scaled the experiment to larger molecules, which made the cover of Nature last month.

Today, we announce the release of OpenFermion, the first open source platform for translating problems in chemistry and materials science into quantum circuits that can be executed on existing platforms. OpenFermion is a library for simulating the systems of interacting electrons (fermions) which give rise to the properties of matter. Prior to OpenFermion, quantum algorithm developers would need to learn a significant amount of chemistry and write a large amount of code hacking apart other codes to put together even the most basic quantum simulations. While the project began at Google, collaborators at ETH Zurich, Lawrence Berkeley National Labs, University of Michigan, Harvard University, Oxford University, Dartmouth University, Rigetti Computing and NASA all contributed to alpha releases. You can learn more details about this release in our paper, OpenFermion: The Electronic Structure Package for Quantum Computers.

One way to think of OpenFermion is as a tool for generating and compiling physics equations which describe chemical and material systems into representations which can be interpreted by a quantum computer1. The most effective quantum algorithms for these problems build upon and extend the power of classical quantum chemistry packages used and developed by research chemists across government, industry and academia. Accordingly, we are also releasing OpenFermion-Psi4 and OpenFermion-PySCF which are plugins for using OpenFermion in conjunction with the classical electronic structure packages Psi4 and PySCF.

The core OpenFermion library is designed in a quantum programming framework agnostic way to ensure compatibility with various platforms being developed by the community. This allows OpenFermion to support external packages which compile quantum assembly language specifications for diverse hardware platforms. We hope this decision will help establish OpenFermion as a community standard for putting quantum chemistry on quantum computers. To see how OpenFermion is used with diverse quantum programming frameworks, take a look at OpenFermion-ProjectQ and Forest-OpenFermion - plugins which link OpenFermion to the externally developed circuit simulation and compilation platforms known as ProjectQ and Forest.

The following workflow describes how a quantum chemist might use OpenFermion in order to simulate the energy surface of a molecule (for instance, by preparing the sort of quantum computation we described in our past blog post):
  1. The researcher initializes an OpenFermion calculation with specification of:
    • An input file specifying the coordinates of the nuclei in the molecule.
    • The basis set (e.g. cc-pVTZ) that should be used to discretize the molecule.
    • The charge and spin multiplicity (if known) of the system.
  1. The researcher uses the OpenFermion-Psi4 plugin or the OpenFermion-PySCF plugin to perform scalable classical computations which are used to optimally stage the quantum computation. For instance, one might perform a classical Hartree-Fock calculation to choose a good initial state for the quantum simulation.
  2. The researcher then specifies which electrons are most interesting to study on a quantum computer (known as an active space) and asks OpenFermion to map the equations for those electrons to a representation suitable for quantum bits, using one of the available procedures in OpenFermion, e.g. the Bravyi-Kitaev transformation.
  3. The researcher selects a quantum algorithm to solve for the properties of interest and uses a quantum compilation framework such as OpenFermion-ProjectQ to output the quantum circuit in assembly language which can be run on a quantum computer. If the researcher has access to a quantum computer, they then execute the experiment.
A few examples of what one might do with OpenFermion are demonstrated in ipython notebooks here, here and here. While quantum simulation is widely recognized as one of the most important applications of quantum computing in the near term, very few quantum computer scientists know quantum chemistry and even fewer chemists know quantum computing. Our hope is that OpenFermion will help to close the gap between these communities and bring the power of quantum computing to chemists and material scientists. If you’re interested, please checkout our GitHub repository - pull requests welcome!


1 If we may be allowed one sentence for the experts: the primary function of OpenFermion is to encode the electronic structure problem in second quantization defined by various basis sets and active spaces and then to transform those operators into spin Hamiltonians using various isomorphisms between qubit and fermion algebras.

TensorFlow Lattice: Flexibility Empowered by Prior Knowledge



(Cross-posted on the Google Open Source Blog)

Machine learning has made huge advances in many applications including natural language processing, computer vision and recommendation systems by capturing complex input/output relationships using highly flexible models. However, a remaining challenge is problems with semantically meaningful inputs that obey known global relationships, like “the estimated time to drive a road goes up if traffic is heavier, and all else is the same.” Flexible models like DNNs and random forests may not learn these relationships, and then may fail to generalize well to examples drawn from a different sampling distribution than the examples the model was trained on.

Today we present TensorFlow Lattice, a set of prebuilt TensorFlow Estimators that are easy to use, and TensorFlow operators to build your own lattice models. Lattices are multi-dimensional interpolated look-up tables (for more details, see [1--5]), similar to the look-up tables in the back of a geometry textbook that approximate a sine function. We take advantage of the look-up table’s structure, which can be keyed by multiple inputs to approximate an arbitrarily flexible relationship, to satisfy monotonic relationships that you specify in order to generalize better. That is, the look-up table values are trained to minimize the loss on the training examples, but in addition, adjacent values in the look-up table are constrained to increase along given directions of the input space, which makes the model outputs increase in those directions. Importantly, because they interpolate between the look-up table values, the lattice models are smooth and the predictions are bounded, which helps to avoid spurious large or small predictions in the testing time.

How Lattice Models Help You
Suppose you are designing a system to recommend nearby coffee shops to a user. You would like the model to learn, “if two cafes are the same, prefer the closer one.” Below we show a flexible model (pink) that accurately fits some training data for users in Tokyo (purple), where there are many coffee shops nearby. The pink flexible model overfits the noisy training examples, and misses the overall trend that a closer cafe is better. If you used this pink model to rank test examples from Texas (blue), where businesses are spread farther out, you would find it acted strangely, sometimes preferring farther cafes!
Slice through a model’s feature space where all the other inputs stay the same and only distance changes. A flexible function (pink) that is accurate on training examples from Tokyo (purple) predicts that a cafe 10km-away is better than the same cafe if it was 5km-away. This problem becomes more evident at test-time if the data distribution has shifted, as shown here with blue examples from Texas where cafes are spread out more.
A monotonic flexible function (green) is both accurate on training examples and can generalize for Texas examples compared to non-monotonic flexible function (pink) from the previous figure.
In contrast, a lattice model, trained over the same example from Tokyo, can be constrained to satisfy such a monotonic relationship and result in a monotonic flexible function (green). The green line also accurately fits the Tokyo training examples, but also generalizes well to Texas, never preferring farther cafes.

In general, you might have many inputs about each cafe, e.g., coffee quality, price, etc. Flexible models have a hard time capturing global relationships of the form, “if all other inputs are equal, nearer is better, ” especially in parts of the feature space where your training data is sparse and noisy. Machine learning models that capture prior knowledge (e.g. how inputs should impact the prediction) work better in practice, and are easier to debug and more interpretable.

Pre-built Estimators
We provide a range of lattice model architectures as TensorFlow Estimators. The simplest estimator we provide is the calibrated linear model, which learns the best 1-d transformation of each feature (using 1-d lattices), and then combines all the calibrated features linearly. This works well if the training dataset is very small, or there are no complex nonlinear input interactions. Another estimator is a calibrated lattice model. This model combines the calibrated features nonlinearly using a two-layer single lattice model, which can represent complex nonlinear interactions in your dataset. The calibrated lattice model is usually a good choice if you have 2-10 features, but for 10 or more features, we expect you will get the best results with an ensemble of calibrated lattices, which you can train using the pre-built ensemble architectures. Monotonic lattice ensembles can achieve 0.3% -- 0.5% accuracy gain compared to Random Forests [4], and these new TensorFlow lattice estimators can achieve 0.1 -- 0.4% accuracy gain compared to the prior state-of-the-art in learning models with monotonicity [5].

Build Your Own
You may want to experiment with deeper lattice networks or research using partial monotonic functions as part of a deep neural network or other TensorFlow architecture. We provide the building blocks: TensorFlow operators for calibrators, lattice interpolation, and monotonicity projections. For example, the figure below shows a 9-layer deep lattice network [5].
Example of a 9-layer deep lattice network architecture [5], alternating layers of linear embeddings and ensembles of lattices with calibrators layers (which act like a sum of ReLU’s in Neural Networks). The blue lines correspond to monotonic inputs, which is preserved layer-by-layer, and hence for the entire model. This and other arbitrary architectures can be constructed with TensorFlow Lattice because each layer is differentiable.
In addition to the choice of model flexibility and standard L1 and L2 regularization, we offer new regularizers with TensorFlow Lattice:
  • Monotonicity constraints [3] on your choice of inputs as described above.
  • Laplacian regularization [3] on the lattices to make the learned function flatter.
  • Torsion regularization [3] to suppress un-necessary nonlinear feature interactions.
We hope TensorFlow Lattice will be useful to the larger community working with meaningful semantic inputs. This is part of a larger research effort on interpretability and controlling machine learning models to satisfy policy goals, and enable practitioners to take advantage of their prior knowledge. We’re excited to share this with all of you. To get started, please check out our GitHub repository and our tutorials, and let us know what you think!

Acknowledgements
Developing and open sourcing TensorFlow Lattice was a huge team effort. We’d like to thank all the people involved: Andrew Cotter, Kevin Canini, David Ding, Mahdi Milani Fard, Yifei Feng, Josh Gordon, Kiril Gorovoy, Clemens Mewald, Taman Narayan, Alexandre Passos, Christine Robson, Serena Wang, Martin Wicke, Jarek Wilkiewicz, Sen Zhao, Tao Zhu

References
[1] Lattice Regression, Eric Garcia, Maya Gupta, Advances in Neural Information Processing Systems (NIPS), 2009
[2] Optimized Regression for Efficient Function Evaluation, Eric Garcia, Raman Arora, Maya R. Gupta, IEEE Transactions on Image Processing, 2012
[3] Monotonic Calibrated Interpolated Look-Up Tables, Maya Gupta, Andrew Cotter, Jan Pfeifer, Konstantin Voevodski, Kevin Canini, Alexander Mangylov, Wojciech Moczydlowski, Alexander van Esbroeck, Journal of Machine Learning Research (JMLR), 2016
[4] Fast and Flexible Monotonic Functions with Ensembles of Lattices, Mahdi Milani Fard, Kevin Canini, Andrew Cotter, Jan Pfeifer, Maya Gupta, Advances in Neural Information Processing Systems (NIPS), 2016
[5] Deep Lattice Networks and Partial Monotonic Functions, Seungil You, David Ding, Kevin Canini, Jan Pfeifer, Maya R. Gupta, Advances in Neural Information Processing Systems (NIPS), 2017


TensorFlow Lattice: Flexibility Empowered by Prior Knowledge

Crossposted on the Google Research Blog

Machine learning has made huge advances in many applications including natural language processing, computer vision and recommendation systems by capturing complex input/output relationships using highly flexible models. However, a remaining challenge is problems with semantically meaningful inputs that obey known global relationships, like “the estimated time to drive a road goes up if traffic is heavier, and all else is the same.” Flexible models like DNNs and random forests may not learn these relationships, and then may fail to generalize well to examples drawn from a different sampling distribution than the examples the model was trained on.

Today we present TensorFlow Lattice, a set of prebuilt TensorFlow Estimators that are easy to use, and TensorFlow operators to build your own lattice models. Lattices are multi-dimensional interpolated look-up tables (for more details, see [1--5]), similar to the look-up tables in the back of a geometry textbook that approximate a sine function.  We take advantage of the look-up table’s structure, which can be keyed by multiple inputs to approximate an arbitrarily flexible relationship, to satisfy monotonic relationships that you specify in order to generalize better. That is, the look-up table values are trained to minimize the loss on the training examples, but in addition, adjacent values in the look-up table are constrained to increase along given directions of the input space, which makes the model outputs increase in those directions. Importantly, because they interpolate between the look-up table values, the lattice models are smooth and the predictions are bounded, which helps to avoid spurious large or small predictions in the testing time.

How Lattice Models Help You

Suppose you are designing a system to recommend nearby coffee shops to a user. You would like the model to learn, “if two cafes are the same, prefer the closer one.”  Below we show a flexible model (pink) that accurately fits some training data for users in Tokyo (purple), where there are many coffee shops nearby.  The pink flexible model overfits the noisy training examples, and misses the overall trend that a closer cafe is better. If you used this pink model to rank test examples from Texas (blue), where businesses are spread farther out, you would find it acted strangely, sometimes preferring farther cafes!
Slice through a model’s feature space where all the other inputs stay the same and only distance changes. A flexible function (pink) that is accurate on training examples from Tokyo (purple) predicts that a cafe 10km-away is better than the same cafe if it was 5km-away. This problem becomes more evident at test-time if the data distribution has shifted, as shown here with blue examples from Texas where cafes are spread out more.
A monotonic flexible function (green) is both accurate on training examples and can generalize for Texas examples compared to non-monotonic flexible function (pink) from the previous figure.

In contrast, a lattice model, trained over the same example from Tokyo, can be constrained to satisfy such a monotonic relationship and result in a monotonic flexible function (green). The green line also accurately fits the Tokyo training examples, but also generalizes well to Texas, never preferring farther cafes.

In general, you might have many inputs about each cafe, e.g., coffee quality, price, etc. Flexible models have a hard time capturing global relationships of the form, “if all other inputs are equal, nearer is better, ” especially in parts of the feature space where your training data is sparse and noisy. Machine learning models that capture prior knowledge (e.g.  how inputs should impact the prediction) work better in practice, and are easier to debug and more interpretable.

Pre-built Estimators

We provide a range of lattice model architectures as TensorFlow Estimators. The simplest estimator we provide is the calibrated linear model, which learns the best 1-d transformation of each feature (using 1-d lattices), and then combines all the calibrated features linearly. This works well if the training dataset is very small, or there are no complex nonlinear input interactions. Another estimator is a calibrated lattice model. This model combines the calibrated features nonlinearly using a two-layer single lattice model, which can represent complex nonlinear interactions in your dataset. The calibrated lattice model is usually a good choice if you have 2-10 features, but for 10 or more features, we expect you will get the best results with an ensemble of calibrated lattices, which you can train using the pre-built ensemble architectures. Monotonic lattice ensembles can achieve 0.3% -- 0.5% accuracy gain compared to Random Forests [4], and these new TensorFlow lattice estimators can achieve 0.1 -- 0.4% accuracy gain compared to the prior state-of-the-art in learning models with monotonicity [5].

Build Your Own

You may want to experiment with deeper lattice networks or research using partial monotonic functions as part of a deep neural network or other TensorFlow architecture. We provide the building blocks: TensorFlow operators for calibrators, lattice interpolation, and monotonicity projections. For example, the figure below shows a 9-layer deep lattice network [5].


Example of a 9-layer deep lattice network architecture [5], alternating layers of linear embeddings and ensembles of lattices with calibrators layers (which act like a sum of ReLU’s in Neural Networks). The blue lines correspond to monotonic inputs, which is preserved layer-by-layer, and hence for the entire model. This and other arbitrary architectures can be constructed with TensorFlow Lattice because each layer is differentiable.

In addition to the choice of model flexibility and standard L1 and L2 regularization, we offer new regularizers with TensorFlow Lattice:
  • Monotonicity constraints [3] on your choice of inputs as described above.
  • Laplacian regularization [3] on the lattices to make the learned function flatter.
  • Torsion regularization [3] to suppress un-necessary nonlinear feature interactions.
We hope TensorFlow Lattice will be useful to the larger community working with meaningful semantic inputs. This is part of a larger research effort on interpretability and controlling machine learning models to satisfy policy goals, and enable practitioners to take advantage of their prior knowledge. We’re excited to share this with all of you. To get started, please check out our GitHub repository and our tutorials, and let us know what you think!

By Maya Gupta, Research Scientist, Jan Pfeifer, Software Engineer and Seungil You, Software Engineer

Acknowledgements

Developing and open sourcing TensorFlow Lattice was a huge team effort. We’d like to thank all the people involved: Andrew Cotter, Kevin Canini, David Ding, Mahdi Milani Fard, Yifei Feng, Josh Gordon, Kiril Gorovoy, Clemens Mewald, Taman Narayan, Alexandre Passos, Christine Robson, Serena Wang, Martin Wicke, Jarek Wilkiewicz, Sen Zhao, Tao Zhu

References

[1] Lattice Regression, Eric Garcia, Maya Gupta, Advances in Neural Information Processing Systems (NIPS), 2009
[2] Optimized Regression for Efficient Function Evaluation, Eric Garcia, Raman Arora, Maya R. Gupta, IEEE Transactions on Image Processing, 2012
[3] Monotonic Calibrated Interpolated Look-Up Tables, Maya Gupta, Andrew Cotter, Jan Pfeifer, Konstantin Voevodski, Kevin Canini, Alexander Mangylov, Wojciech Moczydlowski, Alexander van Esbroeck, Journal of Machine Learning Research (JMLR), 2016
[4] Fast and Flexible Monotonic Functions with Ensembles of Lattices, Mahdi Milani Fard, Kevin Canini, Andrew Cotter, Jan Pfeifer, Maya Gupta, Advances in Neural Information Processing Systems (NIPS), 2016
[5] Deep Lattice Networks and Partial Monotonic Functions, Seungil You, David Ding, Kevin Canini, Jan Pfeifer, Maya R. Gupta, Advances in Neural Information Processing Systems (NIPS), 2017