Roust wants NLP tools performing discovery on data feeding into CFTC‘s enterprise catalog, so unstructured data can be pushed into a structured form when necessary.

CFTC employees need quality, standardized data to effectively oversee product markets and keep the public informed on market activity, work that increasingly involves visualizations and open-data portals — in keeping with the Customer Experience (CX) Executive Order issued in November.

“We have a tremendous amount of futures market experience in this organization, and it is not their subject matter expertise to be trawling through PDF files to find the one nugget of information they’re interested in,” Roust said, during the Data Coalition‘s RegTech22 Data Summit on Wednesday. “That is not the best use of human capital in the government or elsewhere.”

CFTC packages data in products like the weekly Commitments of Traders report it posts on its website for business reporters and institutions, who in turn relay that information to producers.

Increasing visualizations and open-data portals will reduce the amount of time it takes users to consume CFTC information, dubbed the “time tax” in the CX Executive Order, Roust said.

While visualizations will allow users to dive deeper into data on the product markets they’re interested in, there are limits on what CFTC shares publicly because there are privacy requirements around financial institutions.

When CFTC requires more or different data from markets, it writes new rules to obtain it, but the regulator won’t know everything it needs without quality data, Roust said.

Another financial regulation agency, the Consumer Financial Protection Bureau (CFPB), is looking into Platform-as-a-Service technology in order to improve its CX.

Director of Design and Development Adam Scott at the CFPB said his “dream state” is testing new tools and capabilities entirely in the cloud to figure out what works without worrying about IT infrastructure.

The CX Executive Order had a minimal impact on Scott’s team because it was already focused on user-centered design, but its work was legitimized. The executive order’s focus on moments in a person’s life including financing a business, times of crisis like poverty and retirement presents a chance for more interagency CX projects, Scott said.

“I think in the long run where we have a lot of opportunity within government is to see more cross-agency collaboration in some of those focus areas, where we’ve seen that be successful in times of crisis, such as COVID-19,” he said.

The post CFTC eyes natural language processing to standardize data, improve customer experience appeared first on FedScoop.

On top of that, a dozen machine learning and artificial intelligence projects are in the pilot or developmental phase, while four more are fully operational, according to an agency spokesperson.

The projects are part of GSA‘s “eliminate, optimize or automate” effort over the last two years, an effort that’s only speeding up over time, the spokesperson said.

“We expect that the velocity of AI/ML adoption will accelerate similar to our [robotic process automation] program over the next few years,” the spokesperson said. “The various pilots and projects are on different deployment timeframes but cover all our primary mission areas including Public Buildings Service, Federal Acquisition Service, finance, IT and HR.”

One such project is the Solicitation Review, which uses supervised ML to predict whether federal IT solicitations posted to beta.SAM.gov are compliant with Section 508 of the Rehabilitation Act, which lays out IT accessibility requirements. The tool helps GSA employees more efficiently review solicitations and reduces the risk of noncompliance.

A second automation project is a virtual assistant that provides employees with IT self-help capabilities.

GSA doesn’t rely on one procurement method for AI services, instead using a number of contracts to encourage competition and equity among small and disadvantaged businesses. Such contracts are made available to other agencies as well.

Automation spans a number of technologies including ML, natural language processing, chatbots and RPA — the last of which is often the lowest-hanging fruit for agencies. The State Department, Social Security Program, U.S. Patent and Trademark Office, Department of Labor, Army, Air Force, and Navy are among the agencies that have RPA programs.

A big reason automation projects are on the rise governmentwide is the Federal RPA Community of Practice (CoP) and its voluntary leadership team within GSA, said Jim Walker, chief technology officer at UiPath, during a recent ACT-IAC event.

The government-only user group launched in 2019 and has grown to 69 member agencies and about 1,200 attendees on monthly calls.

In November, the CoP issued a State of Federal RPA report —the first detailed review of the technology across government. The report found a 110% increase in deployed automations between fiscal 2019 and 2020.

Additionally, the report found a 195% increase in capacity hours created.

The CoP created a maturity model for agencies to gauge their RPA progress and saw a 70% increase in Level 4 projects, which went from zero to five between fiscal 2019 and 2020.

While only 23 agencies participated in the first report, that number should grow with enthusiasm for RPA.

GSA Chief Financial Officer Gerard Badorrek recently oversaw a 100-day, industry-wide challenge to create RPA solutions that improved the experience for agencies submitting budget justifications. A total of 10 RPA solutions came out of the event, and 12 employees were trained in the technology.

The post GSA leads rise in automation projects governmentwide appeared first on FedScoop.

Natural language processing helps harvest names for adjudicators and flag potential fraud when applicants’ stories don’t align, machine learning (ML) combs biographic and biometric data to identify people with USCIS benefits, and network analytics make connections regarding their relationships and employers, Brown said.

New tools will dissect supporting evidence related to immigration cases, making it easier for adjudicators to make decisions to award people benefits like green cards.

“Now we start to think about a lot of that pre-processing of adjudication really up front, as opposed to it being manually done or swivel chaired at an adjudicator’s workstation or workstations,” Brown said during an AI in Government event. “So providing a lot of that information upfront.”

Computer vision and optical character recognition will be used to validate documents and classify evidence, so adjudicators can click on what they want rather than sort through.

Identity proofing like mobile verification and sentiment analysis are proving more challenging, Brown said.

“We, I feel, need industry experts and assistance in looking at what does this mean from a privacy perspective and abating some of the challenges therein,” he said. “What does this mean from a security perspective?”

Identity validation presents a number of cyberattack vectors when doing something as seemingly benign as verifying photos or videos of people.

Presentation-layer, man-in-the-middle, and backend and data poisoning attacks are all possible.

“Simple things like Avatarify and even TikTok technologies have creeped in,” Brown said. “So I feel this is an area we need a lot of help with.”

Brown also hopes to deal with ML and artificial intelligence “sprawl” by consolidating toolsets and platforms to provide a more robust continuous integration/continuous delivery (CI/CD) pipeline.

Proper experimentation on algorithms that accounts for security and their sharing is also important, Brown said.

USCIS is still trying to solve the problem of data bias by automating algorithms to filter out biased data, audit pipelines and flag where data quality issues persist, Brown said.

Brown hopes to see more adaptive automated services embedding customer and adjudicator personas before 2025.

The post USCIS automating pre-processing of immigration cases appeared first on FedScoop.

So the Navy recently acquired a new supply chain risk assessment tool from KSM Consulting for quicker analysis of its microelectronics and to serve as an example in monitoring broader supply chains for IT-related products.

To determine if a product is safe for the Navy’s networks, analysts pore over documents that show if a company is vulnerable to adversary influence. Much of that requires collecting financial records, public data and private information held by the DOD, a process that on average takes around 80 hours of work — or, “a bunch of googling,” John Roach, executive vice president of data analytics at KSM Consulting, told FedScoop.

“We have automated all of that data gathering process,” Roach said in an interview.

The new tool acquired by Naval Surface Warfare Center-Crane Division — the Automated Microelectronics Analysis & Reporting Optimization (AMARO) solution — uses natural language processing to gather public and private information on companies in the microelectronic supply chain and extract important text. This allows analysts to get straight to the point determining a company’s supply chain trustworthiness.

AMARO is cloud-native and built to be scaled. It also shows a path forward for broader supply chain management tools on other IT-related products the DOD has been looking for.

“We are excited to roll out this technology to the field,” said Adam Hauch, supply chain awareness and security technical lead for DOD. “The AMARO tool will allow us to quickly and thoroughly examine the supply chain of commercial microelectronics, as well as identify vulnerabilities and over-reliance in a more strategic manner.”

Microelectronics are the core of DOD’s vast hardware-based networks and can cause major problems if they are faulty or compromised by adversaries, Roach said.

“At the end of the day everyone thinks about software, but the software also controls some piece of hardware,” he said.

The acquisition comes as senior DOD officials have said they are looking to move to a zero-trust model for microelectronics. That would mean the department assumes a level of risk inherent in the product, requiring security checks at every step of the way instead of a one-and-done approval.

Even in a zero-trust approach, Roach said supply chain monitoring will still be needed for chips and other products. Having better insights into the supply chain will give the Navy and other parts of the DOD more information on a host of challenges, from deciding whether to buy new products or repair existing ones to determining where threats from foreign ownership are.

“This is a capability that will directly lead to improving our supply chain awareness and security,” Hauch said.

The post Navy automates supply chain analysis for microelectronics appeared first on FedScoop.

This marks an evolution for the JAIC as money for battlefield AI is outpacing projects that had an earlier head start, like the center’s continued humanitarian assistance work, JAIC acting Director Nand Mulchandani said Wednesday, without revealing specific budget numbers.

The Warfighting Mission Initiative got its first major contract award in April, with an $800 million deal going to Booz Allen Hamilton to help the JAIC coordinate with other private sector companies on bringing AI to the battlefield. Other mission initiatives include warfighter health, business process transformation, threat reduction and protection, joint logistics, and joint information warfare which includes cyber.

The JAIC has stressed that at every step of development, its ethics principles are considered and inform how the department works to implement AI into kill chains.

“I think about them all the time,” Mulchandani said about the ethics principles at a press conference.

One of the JAIC’s major focuses in its warfighting is developing AI that enables Joint All Domain Command and Control (JADC2), the next-generation network-of-networks the military is developing to link operations across sea, air, land, space and cyber. The AI in JADC2 will be used to link “every sensor to every shooter,” so the saying goes, by having a common data architecture and means to rapidly process that data to get information to the right part of the chain of command.

“That is going to be a big focus,” Mulchandani said of the program.

Another high-priority focus area is in what the JAIC calls “cognitive assistance.” This type of AI technology is designed to filter information and assist service members in the decisionmaking processes. Mulchandani described cognitive assistance as AI that is designed to assist those assessing potential targets of attack. In particular, the JAIC is working with natural language processing (NLP) that can sift through layers of text data to assess the important information and synthesize less valuable text.

Much of the JAIC’s work is happening in partnership with industry, Mulchandani said. Despite protests and worry inside some Silicon Valley companies, he said engagement remains strong both with large companies, like Google and Microsoft, and with small startups.

“We have deep engagement with industry,” he said, adding that the JAIC’s industry engagement specialist is based right in the heart of Silicon Valley.

Mulchandani pointed to offensive information warfare and cyber weapons as an area in which industry hasn’t yet progressed much. While building automated cybersecurity and event analysis tools has been a “well-trodden path” for industry for years, Mulchandani noted that when it comes to offensive cyber tools, “industry has just barely started.”

“There is a huge goldmine of work there,” he said.

Mulchandani said he could not go into details, but noted that JAIC is working with U.S. Cyber Command on some projects around offensive cybersecurity work.

The post The vast majority of JAIC’s money is going toward warfighting appeared first on FedScoop.

As part of what it’s calling the “year of AI delivery,” the Joint Artificial Intelligence Center  intends to field AI-enabled systems under its “Business Process Transformation” initiative. The work isn’t about deploying “killer robots” or whatever else might be in the public imagination, but it is intended to have an immediate impact, according to a blog post Thursday from the JAIC.

The initiative, which has six lines of effort from business administration to human capital management, seeks to leverage different forms of AI for different use cases across the different military services.

“The Business Process Transformation [mission initiative] is developing solutions for digital work streams that realize cost savings and increase productivity at scale,” The JAIC wrote in the blog post. The new systems are a part of the JAIC’s effort to field new tech, not only make policy about it.

Two robotic process automation (RPA)-based challenges the center is working focus on financial management systems and payroll. Pay-change requests remain stuck in backlogged systems the JAIC is looking to modernize. The JAIC wants to use RPAs to automate forms that process the addition of new children and marriage to a sailor or Marine’s personnel file to remove the work from a humans portfolio. For the Army’s financial management system, it needs machine learning to better guide the RPAs already in place.

Working with the Air Force, the AI center is trying to field “logic-based expert systems” to review memos for their compliance with Air Force writing style. Currently, memos need to be reviewed at every level of the chain of command before being advanced, a tedious process of grammatical review that automation could replace. Using AI instead of humans could speed up memo movement by 50 percent, according to the JAIC. Eventually, the center wants to add machine learning to direct where to send memos based on the text it is processing.

For the entire department, the JAIC wants to field natural language processing (NLP) that will search databases of policy memos to ensure new memos are non-redundant and don’t conflict with past policy. The NLP would turn unstructured text data into structured data that could be mined for redundancies or contradictions. The JAIC appears to have enough promise in the product that it is named “Policy Analysis Tool (Game Changer).”

The post AI for everybody: JAIC unveils more initiatives to DOD bolster business systems appeared first on FedScoop.

General Dynamics Information Technology (GDIT) won the one-year contract with four one-year options through USPTO‘s Intelligent Automation and Innovation Support Services blanket purchase agreement.

Tools USPTO plans to pilot under the contract will expedite patent and trademark applications and requests, predict fraudulent transactions and reduce backlogs — cutting costs in the process.

“With this new work, we will bring our AI, ML and robotic process automation expertise to help USPTO develop solutions that accelerate the patent and trademark process to benefit American innovators,” said Christopher Hegedus, a vice president and general manager at GDIT, in the announcement.

USPTO Chief Information Officer Jamie Holcombe said his agency was already developing a new patent examiner search tool leveraging ML algorithms in February. And he expressed interest in applying the technology to trademark image searches and detecting fraudulent imagery.

The pilots stand to create new professional functions within USPTO.

“This contract represents an important step in helping USTPO advance the broader federal government goal to advance the adoption of AI and the way it can be applied to enhance machine potential and expand human capacity,” said Paul Nedzbala, a senior vice president at GDIT, in a statement.

The post USPTO expands use of machine learning with $50M contract appeared first on FedScoop.

That sentiment, told to FedScoop by founder and President of Yext Brian Distelburger, was echoed by a host of tech companies opening their products to local, state and federal agencies alike as governments face the challenges of the novel coronavirus pandemic.

Offers range from assistance with remote working to free cloud storage and services for researchers working on a vaccine for the virus, and they extend far beyond those listed in this report. The government too is responding by opening up data, forming partnerships with tech companies and granting new authorities to work with telecommunications companies.

“This is unprecedented and so the response from the tech sector has been unprecedented,” Nick Sinai, former deputy federal CTO, told FedScoop. Challenges during the pandemic range from agencies like the Department of Veterans Affairs responding to the direct effects of the virus to much of the government faltering under the strain of “maximum telework.”

“Tech CEOs are stepping up and offering pro-bono products, services, and support – they want to do anything they can to help,” Sinai added.

Sinai is a senior advisor to the venture capital firm Insight Partners. Many of the companies on the firm’s portfolio are among those offering their services pro-bono. A list of more than 50 companies has been compiled into a spreadsheet to showcase those offering services to assist with benefit enrolment, remote-working and sharing public information.

Distelburger said Yext is in talks with senior federal agency officials to bring its natural language processing to government websites to better understand the sentiment behind the questions people are searching for on the internet related to the coronavirus and help provide people with the answers based on CDC data. The benefits of the service are twofold: For citizens, Yext offers an easy way to quickly find answers to their questions about COVID-19. For governments, it provides aggregated data about what constituents are asking so leaders can tailor their communications. The company has already provided a similar service to the state of New Jersey.

“I am thrilled we can help Apple stores update their hours for the new iPhone launch,” Distelburger said of the company’s typical work, “but boy, this is way more important.”

Other tech companies are finding their services fit critical needs during the pandemic. Smartsheet, a remote project-management software company, has built a free template to track COVID-19 response efforts for government agencies.

“Smartsheet is all about allowing the citizen to be able to achieve more, to have better insight and move faster,” said Amy Frampton, Smartsheet’s vice president of product marketing. Those qualities “become really critical in a time like this.”

Smartsheet’s government-specific tools are FedRAMP authorized and 1,800 agencies across international, federal, local and other government organizations are using the program. The COVID-19 template has been downloaded 10,000 times, the company said.

Tech giants pitching in too

Titans of the technology industry have pitched in as well. Amazon committed $20 million in cloud computing resources “to accelerate diagnostic research, innovation, and development” for both tracking and mitigation initiatives.

“[O]ne area where we have heard an urgent need is in the research and development of diagnostics, which consist of rapid, accurate detection and testing of COVID-19,” said Teresa Carlson, vice president of worldwide public sector at Amazon Web Services. “Better diagnostics will help accelerate treatment and containment, and in time, shorten the course of this epidemic.”

Microsoft announced it’s offering its Healthcare Bot service powered by Microsoft Azure to help agencies answer basic questions using robotic process automation. The CDC is using the bot to “quickly assess the symptoms and risk factors” for people who may have been exposed and suggest what to do next, according to a Microsoft blog post.

Google said it will donate more than $800 million to frontline organizations and small- and medium-sized businesses. The company also built a portal of coronavirus-related information.

And Apple last week formed a partnership with the federal government to build a website that uses Centers for Disease Control and Prevention data to screen people experiencing coronavirus symptoms. The CDC plans to encourage other companies to build additional COVID-19 tools.

Some restrictions

For many companies, working with state and local governments could be easier than the federal government, which requires FedRAMP authorization to ensure cloud products are secure. The General Services Administration, which operates FedRAMP, did not respond to multiple requests for comment on how it is working with the private sector.

“One of the challenges is there are a lot of offers to help,” Sinai said. “Inside of government, you have to figure out what are the most appropriate set of offers to leverage, and that takes knowledgeable people too.”

There is also the matter of federal acquisition laws and regulations, said Joe Stuntz, head of federal and platform at encryption company Virtru and a former Office of Management and Budget official.

“There are reasons why people can’t just give stuff away,” he said. For instance, if an agency accepts a reduced-price or free service from a company that later wins a lucrative contract, the free offer could be fodder for a protest. “Tech can enable mission, but there are rules in there for a reason,” Stuntz added. Virtru is offering its encryption and privacy technologies for free as well.

One workaround is forming partnerships instead of going through the contracting process. A senior administration official told FedScoop that the White House Coronavirus Task Force, the Federal Emergency Management Agency and CDC used a “partnership” for its work with Apple on its COVID-19 website.

“There was no contract awarded and there was no cost to the taxpayers,” the official said.

For agencies going through normal contracting avenues, there is a need for speed and getting money to contractors. The Air Force is working through its AFventures platform to strengthen its small business and emerging technology supply chain, which its acquisition head Will Roper said remains strong.

A silver lining Sinai sees is that agencies across the government are seeing the returns on cloud migrations and modernizing enterprise services. And in the Air Force, Roper said that the emphasis on becoming a digital Air Force laid the groundwork for the department’s telework success.

“This will actually accelerate the move to [Microsoft] Office 365 in the Department of Defense,” he said.

The post ‘CEOs are stepping up’ free-of-charge. How the tech industry is helping government during COVID-19 appeared first on FedScoop.

The three students from Northern Virginia entered their work in a competition that included software companies like Oracle Healthcare and MyCancerDB. Digital consulting company Composite App took the $20,000 first place prize for its solution — a tool for helping patients stay on track with their care plan —  but the clinical trials team got an honorable mention.

The tech sprint was organized by the VA’s new AI institute, and it focused on partnering with outside organizations and companies interested in applying artificial intelligence tools and techniques to VA data.

The high school team’s members — Shreeja Kikkisetti, Ethan Ocasio and Neeyanth Kopparapu — met as part of the Northern Virginia-based nonprofit Girls Computing League. They were unique in a competition otherwise dominated by adult professionals from software and health care companies.

Their solution, the Clinical Trials Selector, takes lab results, diagnosis and demographic data from the VA and Centers for Medicare and Medicaid Services (CMS) and uses natural language processing as part of its technology to match patients to trials from the National Cancer Institute’s clinical trials database.

The current process for matching patients to trials is a bit more laborious for both the patients themselves and for doctors, Gil Alterovitz, the VA’s director of AI, told FedScoop. It can involve veterans inputting their own data into the clinical trials database — a process that can be confusing and time-consuming.

The Clinical Trials team’s solution, on the other hand, envisions a world where veterans don’t need to have an in-depth understanding of their own medical charts — instead they can log into their personalized VA.gov portal and automatically be matched to potential trials. The application is “functional,” team member Ethan Ocasio told FedScoop, but the team is still working to move it out of a sandbox and make it “production ready” for the VA environment.

Shreeja Kikkisetti and Ocasio told FedScoop that the biggest challenge of the sprint was learning to work with medical informatics. But they both said being a part of the sprint was a “fantastic opportunity” to put their coding skills to a real world test. Ultimately, both said they envision the application being an open source framework available for use by a variety of institutions, not just the VA.

Other companies involved in the sprint include Sanford Imagenetics, which created a product for determining of the relevance of pharmacogenetic testing based on patient characteristics, and LifeOmic, which built a visualization tool for the data sharing platform known as the Veterans Precision Oncology Data Commons.

The VA has become increasingly interested in positioning itself as a federal leader in artificial intelligence research and development. The agency has also recently launched AI projects like an effort to use AI to reduce veterans’ wait times for health appointments, and another to scan medical records and evaluate suicide risk as part of the REACH VET program.

The post VA’s AI Tech Sprint yields a tool for matching patients with clinical trials, and more appeared first on FedScoop.

For the first time, DOT is hiring a chief technology officer at FMCSA to help manage the data that keeps the U.S. transportation system safe.

FMCSA ensures commercial motor vehicle drivers pass an annual physical, around 5.5 million each year, and conducts 3.5 million roadside inspections.

“You’ll see an emphasis on digital and really the user experience going forward because our applications increasingly involve knitting together a state and private sector net around the commercial motor vehicle industry,” Daniel Morgan, chief data officer at DOT, told reporters after an AFCEA event Tuesday.

For instance, a new system slated for 2020 will divert drivers who fail random drug testing at work to substance abuse counseling. DOT is also developing a training provider registry for instructors who teach drivers how to operate commercial motor vehicles to ensure a baseline curriculum.

Both are separate websites built one regulation at a time.

“That’s not necessarily the best way to build that user experience,” Morgan said. “So we need somebody to lead that organization.”

But DOT regulations also need to be simplified more broadly on everything from vehicle safety requirements to the proper transportation of hazardous materials, he added. And the department is using natural language processing to do so.

Through a partnership with Amazon Web Services, and using tax increment financing dollars and open-source software from the Mercatus Center at George Mason University, DOT developed a regulatory dashboard.

“We try to figure out ways to help our lawyers target where they can spend their energy for the most value,” Morgan said during a panel discussion. “Ultimately we want more people to comply [with regulations].”

DOT’s tool creates statistics for its lawyers by analyzing Federal Register data to generate metrics like sentence length, word count, reading level and Shannon’s Entropy — a measure of how hard it is for autocorrect to figure out the next word in a sentence.

“A score of 8 or 9 is like Shakespearean, and our regulations score more complex than Shakespeare,” Morgan told reporters.

DOT also measures the number of conditions in its regulations — “must” and “shall” — and exceptions — “but,” “notwithstanding” and “except” — which increase complexity. If the department notices “exception drift,” its lawyers can refactor regulations, Morgan said.

Other agencies with industry-focused regulations could use DOT’s tool, though departments with broader regulations like Agriculture, Health and Human Services, and Labor may have a more difficult time, he added.

The post DOT agencies adding data leadership while simplifying regulations appeared first on FedScoop.