You are here

Feed aggregator

One simple graphic: Researchers love PyTorch and TensorFlow

O'Reilly Radar - Thu, 2019/07/25 - 04:00

Interest in PyTorch among researchers is growing rapidly.

In a recent survey—AI Adoption in the Enterprise, which drew more than 1,300 respondents—we found significant usage of several machine learning (ML) libraries and frameworks. About half indicated they used TensorFlow or scikit-learn, and a third reported they were using PyTorch or Keras.

I recently attended an interesting RISELab presentation delivered by Caroline Lemieux describing recent work on AutoPandas and automation tools that rely on program synthesis. In the course of her presentation, Lemieux reviewed usage statistics they had gathered on different deep learning frameworks and data science libraries. She kindly shared some of that data with me, which I used to draw this chart:

Figure 1. Number of papers posted on arXiv.org that mention each framework. Source: Data from RISELab and graphic by Ben Lorica.

The numbers are based on simple full-text searches of papers posted on the popular e-print service arXiv.org. Specifically, they reflect the number of papers which mention (in a full-text search) each of the frameworks. Using this metric, the two most popular deep learning frameworks among researchers are TensorFlow and PyTorch. From January to the end of June 2019, about 1,800 papers mentioned TensorFlow and a comparable number mentioned PyTorch. Most notably, interest in PyTorch among researchers is growing rapidly: it grew 194% year-over-year (Jan-Jun 2018 compared to Jan-Jun 2019).

To the extent that researchers and teachers are harbingers and strongly influence what future professionals might use, look for PyTorch to also gain users among data scientists, developers, ML engineers, and companies. In a recent post, we outlined the suite of tools for managing machine learning in the enterprise. Many of the tools and companies we highlighted in that post support the popular ML libraries and deep learning frameworks (particularly TensorFlow and PyTorch), thus we predict both of these frameworks will be equally viable options for enterprise users.

Related content:

Continue reading One simple graphic: Researchers love PyTorch and TensorFlow.

Categories: Technology

Four short links: 24 July 2019

O'Reilly Radar - Wed, 2019/07/24 - 05:20

Computer Life, Quantum Hype, Python Anti-patterns, and Algorithm Series

  1. Nils Barricelli (Nautilus) -- history of an artificial (computer) life pioneer.
  2. Quantum Hype (ComputerWorld) -- the quantum computing memes are ace, but so is the general discussion of why quantum computing is felt by insiders to be overhyped.
  3. Python Anti-patterns -- readable collection of things Not To Do.
  4. All Hail the Algorithm -- Al-Jazeera five-part series exploring the impact of algorithms on our everyday lives.

Continue reading Four short links: 24 July 2019.

Categories: Technology

Four short links: 23 July 2019

O'Reilly Radar - Tue, 2019/07/23 - 05:00

Deciphering Linear B, Data Journalism, Innovation Contradictions, Rebuilding Slack

  1. Applying Deep Learning to Linear B -- To compensate for the lack of strong supervision signal, our model design is informed by patterns in language change documented in historical linguistics. [...] When applied to the decipherment of Ugaritic, we achieve a 5.5% absolute improvement over state-of-the-art results. We also report the first automatic results in deciphering Linear B, a syllabic language related to ancient Greek, where our model correctly translates 67.3% of cognates.
  2. Data Science Behind Data Journalism (Chris Knox) -- discusses the data analysis that went into a story on vaccination in NZ. A good example of how to use data to do to journalism (and not just torture it to say what you want).
  3. The Hard Truth About Innovative Cultures (HBR) -- A tolerance for failure requires an intolerance for incompetence. A willingness to experiment requires rigorous discipline. Psychological safety requires comfort with brutal candor. Collaboration must be balanced with individual accountability. And flatness requires strong leadership. Innovative cultures are paradoxical. Unless the tensions created by this paradox are carefully managed, attempts to create an innovative culture will fail. (via Tim Kong)
  4. When a Rewrite Isn't: Rebuilding Desktop Slack -- Our plan was to: keep the existing codebase; create a “modern” section of the codebase that would be future-proof and work the way we wanted it to; modernize the implementation of Slack bit by bit, replacing existing code with modern code incrementally; define rules that would enforce a strict interface between existing and modern code so it would be easy to understand their relationship; and continually ship all of the above with the existing app, replacing older modules with modern implementations that suited our new architecture. The final step—and the most important one for our purposes—was to create a modern-only version of Slack that would start out incomplete but gradually work its way toward feature completeness as modules and interfaces were modernized.

Continue reading Four short links: 23 July 2019.

Categories: Technology

Four short links: 22 July 2019

O'Reilly Radar - Mon, 2019/07/22 - 07:30

Game Source, Procurement Graph, Data Moats, and Antitrust Regulation

  1. Game Source Code -- Internet Archive has a collection of video game source code. The majority of these titles were originally released as commercial products and the source code was made available to the public at a later time.
  2. European Public Procurement Knowledge Graph -- over 23 million triples (records), covering information about almost 220,000 tenders, built to support competitiveness and accountability by TheyBuyForYou. (via University of Southampton)
  3. The Empty Promise of Data Moats (Andreessen-Horowitz) -- business model wonks reckoned that "data network effects" were a thing, but the benefits seen by companies claiming data network effects seem to be the benefits of simply having a lot of data. And that's not as defensible as hoped. I liked this essay.
  4. Why Big Tech Keeps Outsmarting Antitrust Regulators (Tim O'Reilly) -- designers of marketplace-platform algorithms and screen layouts can arbitrarily allocate value to whom they choose. The marketplace is designed and controlled by its owners, and that design shapes “who gets what and why.” [...] Power over sellers ultimately translates into power over customers as well. When it comes to antitrust, the question of market power must be answered by analyzing the effect of these marketplace designs on both buyers and sellers, and how they change over time. How much of the value goes to the platform, how much to consumers, and how much to suppliers?

Continue reading Four short links: 22 July 2019.

Categories: Technology

Four short links: 19 July 2019

O'Reilly Radar - Fri, 2019/07/19 - 10:05

Journal Mining, API Use, Better Conversation, and Apollo 11 Source

  1. 73 Million Journal Articles for Text Mining (BoingBoing) -- The JNU Data Depot is a joint project between rogue archivist Carl Malamud, bioinformatician Andrew Lynn, and a research team from New Delhi's Jawaharlal Nehru University: together, they have assembled 73 million journal articles from 1847 to the present day and put them into an airgapped respository that they're offering to noncommercial third parties who want to perform textual analysis on them to "pull out insights without actually reading the text."
  2. How Developers Use API Documentation: An Observation Study (ACM) -- participants totally mapped to opportunistic (risk-taking, paste-then-adapt, change-without-checking) developers and systematic (start with clean code, read the docs, learn before coding) developers.
  3. Talk -- An open source commenting platform focused on better conversation.
  4. Apollo 11 -- Original Apollo 11 Guidance Computer (AGC) source code for the command and lunar modules.

Continue reading Four short links: 19 July 2019.

Categories: Technology

O'Reilly Open Source and Frank Willison Awards

O'Reilly Radar - Thu, 2019/07/18 - 13:00

The O’Reilly Open Source Awards recognize individual contributors who have demonstrated exceptional leadership, creativity, and collaboration in the development of open source software.

Continue reading O'Reilly Open Source and Frank Willison Awards.

Categories: Technology

The war for the soul of open source

O'Reilly Radar - Thu, 2019/07/18 - 13:00

Drawing on 13 years spent building the Chef community, Adam Jacob takes a deep dive into the soul of open source.

Continue reading The war for the soul of open source.

Categories: Technology

Open source force multipliers

O'Reilly Radar - Thu, 2019/07/18 - 13:00

Adrian Cockcroft says the most successful open-source-based businesses have turned their partners and developer communities into force multipliers for their own marketing and engineering teams.

Continue reading Open source force multipliers.

Categories: Technology

O’Reilly Radar: Open source technology trends—What our users tell us

O'Reilly Radar - Thu, 2019/07/18 - 13:00

Using aggregate analysis of O’Reilly online learning content usage and search data, Roger Magoulas shares key insights that impact the technology tools ecosystem.

Continue reading O’Reilly Radar: Open source technology trends—What our users tell us.

Categories: Technology

Ask not what Brands™ can do for you

O'Reilly Radar - Thu, 2019/07/18 - 13:00

VM Brasseur discusses the help new companies need to become authentic members of the free and open source software community.

Continue reading Ask not what Brands™ can do for you.

Categories: Technology

Managing machines

O'Reilly Radar - Thu, 2019/07/18 - 13:00

Pete Skomoroch covers what you need to know as we shift from a world of deterministic programs to systems that give unpredictable results on ever-changing training data.

Continue reading Managing machines.

Categories: Technology

Acquiring and sharing high-quality data

O'Reilly Radar - Thu, 2019/07/18 - 06:30

The O’Reilly Data Show Podcast: Roger Chen on the fair value and decentralized governance of data.

In this episode of the Data Show, I spoke with Roger Chen, co-founder and CEO of Computable Labs, a startup focused on building tools for the creation of data networks and data exchanges. Chen has also served as co-chair of O'Reilly's Artificial Intelligence Conference since its inception in 2016. This conversation took place the day after Chen and his collaborators released an interesting new white paper, Fair value and decentralized governance of data. Current-generation AI and machine learning technologies rely on large amounts of data, and to the extent they can use their large user bases to create “data silos,” large companies in large countries (like the U.S. and China) enjoy a competitive advantage. With that said, we are awash in articles about the dangers posed by these data silos. Privacy and security, disinformation, bias, and a lack of transparency and control are just some of the issues that have plagued the perceived owners of “data monopolies.”

In recent years, researchers and practitioners have begun building tools focused on helping organizations acquire, build, and share high-quality data. Chen and his collaborators are doing some of the most interesting work in this space, and I recommend their new white paper and accompanying open source projects.

Continue reading Acquiring and sharing high-quality data.

Categories: Technology

Four short links: 18 July 2019

O'Reilly Radar - Thu, 2019/07/18 - 01:00

Weird Algorithms, Open Syllabi, Conversational AI, and Quantum Computing

  1. 30 Weird Chess Algorithms (YouTube) — An intricate and lengthy account of several different computer chess topics from my SIGBOVIK 2019 papers. We conduct a tournament of fools with a pile of different weird chess algorithms, ostensibly to quantify how well my other weird program to play color- and piece-blind chess performs. On the way we "learn" about mirrors, arithmetic encoding, perversions of game tree search, spicy oils, and hats.
  2. Open Syllabus Project — as FastCompany explains, the 6M+ syllabi from courses around the world tell us about changing trends in subjects. Not sure how I feel that four of the textbooks I learned on are still in the top 20 (Cormen, Tanenbaum, Silberschatz, Stallings).
  3. Plato — Uber open-sourced its flexible platform for developing conversational AI agents. See also their blog post.
  4. Speediest Quantum Operation Yet (ScienceDaily) — In Professor Michelle Simmons' approach, quantum bits (or qubits) are made from electrons hosted on phosphorus atoms in silicon.[...] "Atom qubits hold the world record for the longest coherence times of a qubit in silicon with the highest fidelities," she says. "Using our unique fabrication technologies, we have already demonstrated the ability to read and initialise single electron spins on atom qubits in silicon with very high accuracy. We've also demonstrated that our atomic-scale circuitry has the lowest electrical noise of any system yet devised to connect to a semiconductor qubit." [...] A two-qubit gate is the central building block of any quantum computer -- and the UNSW team's version of it is the fastest that's ever been demonstrated in silicon, completing an operation in 0.8 nanoseconds, which is ~200 times faster than other existing spin-based two-qubit gates.

Continue reading Four short links: 18 July 2019.

Categories: Technology

The next age of open innovation

O'Reilly Radar - Wed, 2019/07/17 - 13:00

Alison McCauley looks at how blockchain technology offers new tools that can help extend the ethos of open innovation into new areas.

Continue reading The next age of open innovation.

Categories: Technology

Why Amazon cares about open source

O'Reilly Radar - Wed, 2019/07/17 - 13:00

Arun Gupta discusses the reasons why AWS is committed to open projects and communities.

Continue reading Why Amazon cares about open source.

Categories: Technology

Better living through software

O'Reilly Radar - Wed, 2019/07/17 - 13:00

Tiffani Bell shares three lessons she's learned exploring how technology can help the less fortunate.

Continue reading Better living through software.

Categories: Technology

Highlights from the O'Reilly Open Source Software Conference in Portland 2019

O'Reilly Radar - Wed, 2019/07/17 - 13:00

Experts explore the role open source software plays in fields as varied as machine learning, blockchain, disaster response, and more.

People from across the open source world came together in Portland, Ore. for the O'Reilly Open Source Software Conference (OSCON). Below you'll find links to highlights from the event.

The war for the soul of open source

Drawing on 13 years spent building the Chef community, Adam Jacob takes a deep dive into the soul of open source.

Ask not what Brands™ can do for you

VM Brasseur discusses the help new companies need to become authentic members of the free and open source software community.

Better living through software

Tiffani Bell shares three lessons she's learned exploring how technology can help the less fortunate.

O’Reilly Radar: Open source technology trends—What our users tell us

Using aggregate analysis of O’Reilly online learning content usage and search data, Roger Magoulas shares key insights that impact the technology tools ecosystem.

The next age of open innovation

Alison McCauley looks at how blockchain technology offers new tools that can help extend the ethos of open innovation into new areas.

Managing machines

Pete Skomoroch covers what you need to know as we shift from a world of deterministic programs to systems that give unpredictable results on ever-changing training data.

Open source force multipliers

Adrian Cockcroft says the most successful open-source-based businesses have turned their partners and developer communities into force multipliers for their own marketing and engineering teams.

Built to last: Building and growing open source communities

Kay Williams explores key lessons for building strong open source communities based on Microsoft’s real-world experience with Kubernetes and VSCode.

The role of open source in mitigating natural disasters

Pedro Cruz and Brad Topol discuss Call for Code, a global developer competition that uses open source technologies to address natural disasters.

Why Amazon cares about open source

Arun Gupta discusses the reasons why AWS is committed to open projects and communities.

O'Reilly Open Source and Frank Willison Awards

The O’Reilly Open Source Awards recognize individual contributors who have demonstrated exceptional leadership, creativity, and collaboration in the development of open source software.

Sustainable gardens

Using examples from Google-supported open source communities, Paris Pittman offers practical advice for ensuring your community thrives.

Be a docs star

Megan Byrd-Sanicki explores how documentation is the super power your open source project needs to grow adoption.

-->

Continue reading Highlights from the O'Reilly Open Source Software Conference in Portland 2019.

Categories: Technology

The role of open source in mitigating natural disasters

O'Reilly Radar - Wed, 2019/07/17 - 13:00

Pedro Cruz and Brad Topol discuss Call for Code, a global developer competition that uses open source technologies to address natural disasters.

Continue reading The role of open source in mitigating natural disasters.

Categories: Technology

Built to last: Building and growing open source communities

O'Reilly Radar - Wed, 2019/07/17 - 13:00

Kay Williams explores key lessons for building strong open source communities based on Microsoft’s real-world experience with Kubernetes and VSCode.

Continue reading Built to last: Building and growing open source communities.

Categories: Technology

Four short links: 17 July 2019

O'Reilly Radar - Wed, 2019/07/17 - 01:00

Margaret Hamilton, WeChat Censorship, Refactoring, and Ancient Games

  1. Margaret Hamilton Interview (The Guardian) — I found a job to support our family at the nearby Massachusetts Institute of Technology (MIT). It was in the laboratory of Prof Edward Lorenz, the father of chaos theory, working on a system to predict weather. He was asking for math majors. To take care of our daughter, we hired a babysitter. Here I learned what a computer was and how to write software. Computer science and software engineering were not yet disciplines; instead, programmers learned on the job. Lorenz’s love for software experimentation was contagious, and I caught the bug.
  2. How WeChat Censors Images in Private Chats (BoingBoing) — WeChat maintains a massive index of the MD5 hashes of every image that Chinese censors have prohibited. When a user sends another user an image that matches one of these hashes, it's recognized and blocked at the server before it is transmitted to the recipient, with neither the recipient or the sender being informed that the censorship has taken place. Separately, all images not recognized in the hash database are processed out-of-band.
  3. The Best Refactoring You've Never Heard Of (James Koppel) — lambdas vs data structures. Very interesting talk.
  4. Machine Learning is About to Revolutionize the Study of Ancient Games (MIT TR) — The team model games as mathematical entities that lend themselves to computational study. This is based on the idea that games are composed of units of information called ludemes, such as a throw of the dice or the distinctively skewed shape of a knight’s move in chess. Ludemes are equivalent to genes in living things or memes as elements of cultural inheritance. They can be transmitted from one game to another, or they may die, never to be seen again. But a key is that they can be combined into bigger edifices that form games themselves.

Continue reading Four short links: 17 July 2019.

Categories: Technology

Pages

Subscribe to LuftHans aggregator