Data science can be utilized as a solution for many sectors and industries—this was the underlying messaging during the Global Women in Data Science (WiDS) one-day annual technical conference by Obsidian Security, hosted at Stanford University on March 5th, and live streamed at the Cove @ UCI in addition to more than 100 worldwide locations. The conference discussed the latest data science research in several domains, ranging from investments to healthcare and education, and also highlighted certain cutting-edge companies that are leveraging data science to succeed.
Leda Braga, chief executive officer, Systematica Investments, discussed the reality of investment management, how recent developments in data and artificial intelligence (AI) shape the fund management industry and the challenges of dealing with financial data.
“By diversifying the choice of trade opportunity, the systematic approach makes the investment process less reliant on the random nature of forecasting and more reliant on portfolio allocation and risk control,” said Braga during the presentation.
Braga emphasized how the systematic approach to investments will become the majority of future asset management as opposed to discretionary investing. By using data science during the investment process, driven namely by signal generation, portfolio construction using estimation and optimization techniques, the systematic approach to investments proves to be less volatile.
During the presentation’s first Technical Vision Talks segment, Mala Analand, executive vice president, SAP Leonardo Data Analytics, discussed the digitalization of the healthcare industry and how this can improve patient outcomes.
“The data is in silos through the entire value chain of healthcare providers and this limits collaboration,” said Analand. “You need to treat data as a first-class data asset, from there, establish a data framework, so it’s formally and securely managed across the network of data. Then connect the data and make it available through the organization to encourage a culture of data-driven decisions.”
Her company first used visual exploration of clinical and genomic datasets followed by real world evidence to power an improved treatment selection, then data integration to allow clinicians to make an ideal treatment decision and lastly, integrated a large cohort of patient datasets for future outcomes.
Lada Adamic, research scientist manager, Facebook, also spoke about several studies of information diffusion in social networks. She emphasized how cascades, processes that create and execute complex data processing workflows, can be used to predict future data.
The second segment of Technical Vision Talks featured Nathalie Riche, researcher at Microsoft Research, where she presented the latest in research efforts in information visualization and data-driven storytelling. She also discussed research on visual stories and their compelling qualities.
Daniela Witten, associate professor of statistics and biostatics, University of Washington, also talked about the statistical challenges brought on by big data, namely in the biomedical research field, and ways to statistically solve those problems.
Featured keynote address speaker Latanya Sweeney, Ph.D., professor of government and technology in residence, Harvard University, discussed data security. Sweeney uses technology to assess and solve societal, governmental and political problems.
“We live in a world in which technology design dictates the rules that we live by,” said Dr. Sweeney. “Everything from location tracking to the price of a ride to self-driving cars, fraud prevention to face recognition, we live in a technocracy and unforeseen consequences of this [AI] technology continue to be seen today.”
During her presentation, she emphasized how data can be utilized as a form of invasion and discrimination, and cited an example of googling her own name only to see an ad for arrest records, of which she had none. Throughout the presentation, she showed the audience how simple data information, such as first names, are utilized as a form of racial discrimination through specific algorithmic bias. By using a virtual private network (VPN), Dr. Sweeney collected 140,000 ad deliveries when searching for the names of actual Americans.
“We found that if you had a name more often [given] to black babies than white babies, [than] you were 80 percent more likely to get an arrest ad,” said Sweeney. “This turns out to be a violation of the [U.S.] Civil Rights Act.”
Dr. Sweeney described the disadvantages for employment opportunities because of this type of scenario. She positioned the audience to consider how society thinks of our current laws inside this technical framework.
During the afternoon Technical Vision Talks, keynote speaker Jia Li, head of Cloud R&D, Could AI at Google, discussed the benefits the cloud has for the medical community for efficiency and accuracy. She spoke on how industries should harness the power of AI to collect and interpret data, specifically citing thoracic disease and chest X-rays being a significant problem. Her hope is to use AI-assisted radiology image analysis.
“There is still much more to be explored and innovated in this space,” said Li. “Hopefully in the future, our [medical] specialists can spend less time on repetitive and error-prone tasks by working together on AI assisted tools.”
The top three winners of the WiDS Datathon, a predictive analysis challenge, were recognized and awarded cash prizes. More than 230 teams analyzed a dataset that contained demographic and behavioral information from a representative sample of survey respondents from India and their usage of traditional financial and mobile financial services. Participants also built machine learning and statistical models to predict the gender of each survey respondent.
For more information about WiDS Conference, click here.