Predictive APIs and apps conference series PAPIs.io will host a two-day conference Aug. 6-7 in Sydney. The event covers tutorials, research, predictive tools and Machine Learning APIs, as well as use cases from businesses applying predictive intelligence.
Leading API industry thinkers such as Mark O’Neill from Axway see Australia as a hotbed for API activity. In February, he told ProgrammableWeb that Australian businesses understand the importance of leveraging APIs when addressing a global market. "For businesses to reach suppliers, channels, markets globally, this feeds into their understanding that they need APIs,” O’Neill said at that time. “These customers are not just selling to local markets, and that means connecting with resellers, suppliers and channels all over the world, at any time."
Now, PAPIs.io conference organizer Louis Dorard sees the Australian environment taking a similar interest in working with predictive APIs. Dorard explains:
First of all, Australia has had a very strong machine learning community for many years, with world-renowned research institutions such as NICTA and researchers who made major contributions to the field, such as Ross Quinlan, who invented decision tree learning algorithms (that are among the most widely applicable ML algorithms).
Industry-wise, there’s proof that Australian businesses are hungry for machine learning tools with major events such as the AWS Summit last May that featured Amazon Machine Learning, and with the leading data science conference KDD taking place in Sydney a few days after PAPIs ’15. Also, it was an Australian company, GCS Agile, who was the first to reach out to us last year about the 2015 edition of PAPIs, and they are now a co-organizer.
So far, program speakers include:
- Alex Housley, CEO at Seldon
- Mark Reid, machine learning researcher at ANU/NICTA
- Poul E.J. Petersen, chief infrastructure officer at BigML
- Danny Lange, general manager at Amazon Machine Learning
A preliminary program is published on Lanyrd. The event will be held at the Menzies Hotel in Sydney’s CBD Center.
“Part of PAPIs’ mission is to contribute to defining the future of predictive APIs, so for the first time this year we are introducing a new research track that focuses on technical advances and challenges in the predictive API space,” says Dorard.
It’s worth noting that even though these technologies are production ready, the space is still very new. PAPIs ’15 will be the first event to have all the leading players (Amazon ML, BigML, Google Prediction API, Microsoft Azure ML) present on stage and participate in a panel discussion. What I see is that they have the same vision of making machine learning more accessible, but they each have developed their own solutions in different ways, which does not make the life of the user as easy as it should be (it’s difficult, for instance, to compare the APIs and their performances). One of the talks in the research track will present the first standardization efforts in the form of a specification for delivering flexible ML services via RESTful APIs.
Dorard expects the audience at PAPIs to continue to be as mixed as at previous events. Those interested in predictive algorithms from across marketing, banking and finance, app development, and startups are expected to attend. Dorard says that startups in particular “now have the means to do what used to require large teams of Ph.D.s (such as churn prediction, user-item recommendations, etc.).” He adds: “Large enterprises are also learning about new tools that are making their teams more efficient.”
While there is diversity in the range of industry sectors taking up the potential of predictive analytics, there is one area that is most troubling among the PAPIs events and the data science sector to date: It is fast becoming one of the most gender-unequal subsectors of an already inequitable technology sector overall. All speakers at both this event (so far) and the previous PAPIs Connect in Paris are male. The 21 people who have registered on the Lanyrd site to say they will attend are male. A study last year by O’Reilly found that among the data sciences, women are paid US$13,000 less in the sector, based purely on their gender. Events where women are invisible can contribute to this inequity and reinforce the status quo.
Dorard acknowledges the issue:
At PAPIs ’14 and PAPIs Connect, I met women who were eager to find out more about the predictive space, but there were only two female speakers and there will only be one at PAPIs ’15, so gender inequity is definitely an issue. Most of our talks come from our call for proposals, and they are selected after rigorous review by our program committee. We individually invited women to submit proposals, but this was not enough as there was a very small proportion of proposals coming from women.
In the future, we need to promote our call for papers to communities of women and to show them the variety of topics of interest at PAPIs, going from use cases that discuss the application of predictive technology in specific domains to technical and research papers. Another thing where we can improve is having more women in our program committee — and asking for their help in promoting the call for papers.
It is heartening that Dorard sees that it is an issue and that greater active participation in decision-making is one way to work toward greater inclusion for the sector. Other API conferences have worked hard at trying to overcome this disparity. API Strategy & Practice and APIDays, for example, make clear statements about participant code of conduct at the start of proceedings and actively seek out female and culturally diverse speakers for their conferences rather than just relying on the call for papers, which will tend to reflect any disparities that exist in the sector.
O’Reilly’s SolidCon, which was held recently in San Francisco, opened a number of scholarships to alum of Women Who Code, which created a much more equal level of participation among attendees (and a more vibrant feel to the conference). Solid also included female keynote and session speakers throughout. Since Solid was a conference about hardware, it was promising to see that physical and engineering projects were not solely the domain of men at the event.
With women like Michele Chambers, president and COO of RapidMiner; Laurie Skelly at Datascope; Alice Zheng, director of data science at Dato; and Tanya Berger-Wolf, a leading researcher at University of Illinois, it should be possible to more easily highlight women's contributions to the predictive analytics and machine learning field. One place to start to better understand women’s participation in data science and predictive analytics technologies would be to review the experiences documented in the free e-book by Cornelia Lévy-Bencheton and Shannon Cutt, Women in Data.
More details about PAPIs.io in Sydney are available at the website. Those interested in attending can sign up for the mailing list or register directly.