Uses Archives - Business Intelligence Info

Tag Archives: Uses

Insightin Health raises $12 million for AI that uses big data to guide patients’ decisions

February 10, 2021 Big Data

Insightin Health, a company delivering personalized health care guidance, today announced it has raised $ 12 million. A spokesperson told VentureBeat the round will support Insightin Health’s plan to bring on more plan providers that service primarily Medicare, Medicaid, and accountable care organization members.

The global big data analytics market for health care was valued at $ 16.87 billion in 2017 and is projected to reach $ 67.82 billion by 2025, according to a recent report from Allied Market Research. It’s believed that health care organizations’ implementation of big data analytics might lead to an over 25% reduction in annual costs in the coming years. Better diagnosis and disease predictions, enabled by AI and analytics, can lead to cost reduction by decreasing hospital readmission rates, among other factors.

Baltimore, Maryland-based Insightin Health, which was founded in 2016 by Enam Noor, is a cloud-based marketing platform that allows organizations to develop data hubs that promote health plan sign-ups and retention. Prior to founding Insightin Health, Noor cofounded Insightin Technology, a cloud computing and enterprise software development company, and Desme, an interactive marketing company.

Insightin Health’s campaign automation system can create personalized and rule-based content delivery in real time, with a machine learning- and AI-driven approach that provides recommendations for touchpoints of member communications and activities. The company claims to combine medical, clinical, cognitive, and social determinants of health to recommend the next best action for each health plan member. For example, the platform can encourage sick members to make outbound calls and schedule telehealth or in-person appointments when available.

Insightin Health can also identify social needs at both a member and population level and offer a dashboard visually detailing members’ responses, activities, and challenges. The company cites research from startup HealthMine that found 60% of surveyed Medicare Advantage members say their health plan doesn’t encourage actions to improve health.

“Our proprietary platform and world-class team have put us in a position to transform the way health care is experienced. The continued interest from our investors and customers is a great validation [of] our company’s momentum since our last raise in 2018,” Noor told VentureBeat via email. “With the continued investment we’re putting into our platform, we’re hoping that more and more health insurance providers will be able to communicate with consumers on a one-to-one level. This way, we can get closer to humanizing the health care experience to improve satisfaction, increase engagement, and ultimately improve health.”

Late last year, Insightin Health teamed up with students from the University of Pittsburgh’s Swanson School of Engineering to predict the start of flu season, with the goal of optimizing flu shot timing and reducing health care costs in the process. More recently, Insightin Health launched a solution to support COVID-19 preparedness efforts. It scrubs infection and mortality data, creating a risk scoring model to assess information at a state and county level across the U.S.

According to Noor, Insightin Health has been profitable since launch and experienced 170% year-over-year revenue growth from 2019 to 2020. With over 5 million people on its platform, the company anticipates similar growth from 2020 to 2021.

Blue Venture Fund and Blue Heron Capital co-led the series A round Insightin Health announced today. It brings the company’s total raised to over $ 13 million. Insightin has 45 employees.

VentureBeat

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact.

Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:

up-to-date information on the subjects of interest to you
our newsletters
gated thought-leader content and discounted access to our prized events, such as Transform
networking features, and more

Become a member

Let’s block ads! (Why?)

Big Data – VentureBeat

How Ubiq Security uses APIs to simplify data protection

November 28, 2020 Big Data

As cyberthreats continue to multiply, startups with tools to protect data are in high demand. But companies are now facing the growing complexity of managing security across their various data sources.

San Diego-based Ubiq Security believes APIs could play a key role in simplifying this task. The company hopes to encourage more developers and enterprises to build security directly into applications rather than looking for other services to plug the holes.

“How do you take the messy and complicated world of encryption and distill it down to a consumable, bite-sized chunk?” Ubiq CEO Wias Issa asked. “We built an entirely API-based platform that enables any developer of any skill set to be able to integrate encryption directly into an application without having any prior cryptography experience.”

Issa is a security veteran and said companies have generally been focused on security for their data storage systems. When they start layering applications on top, many developers find they haven’t built security into those products. In addition, the underlying storage is becoming a thicket of legacy and cloud-based solutions.

“You could have an Oracle database, an SQL Server, AWS storage, and then a Snowflake data warehouse,” Issa said. “You’ve got to go buy five or six different tools to do encryption on each one of those because they’re all structured differently.”

Even when encryption is included in the application, it can be poorly designed. Issa said cryptographic errors have typically been among the top three vulnerabilities in software applications over the past decade.

“When you’re a developer in 2020, you’re expected to know multiple languages, do front end, back end, full-stack development,” Issa said. “And on top of that, someone comes along and says, ‘Hey, can you do cryptography?’ And so the developer thinks, ‘How do I just get past this so I can go back to building a fantastic product and focusing on my day job?’ So key management is an area where developers either don’t understand it or don’t want to deal with it because it’s so complicated and so burdensome and, frankly, it’s very expensive to do.”

To cut through those challenges, Ubiq’s API-based developer platform lets developers simply include three lines of code that make two API calls. By handling encryption at the application layer with an API, the security works across all underlying storage systems as well.

“The application will handle all the encryption and decryption and simply hand the data in an encrypted state to the storage layer,” Issa said. “That allows them to not only have a better security posture but improve their threat model and reduce the overall time it takes to roll out an encryption plan.”

Customers can then use a dashboard to monitor their encryption and adjust policies without having to update code or even know the developer jargon. This, in turn, simplifies the management of encryption keys.

Lessons from the government

Among its more notable customers, Ubiq announced this year that it had signed deals with the United States Army and the U.S. Department of Homeland Security. While government buyers have their particular issues, in this case the military and civilian systems faced many of the same obstacles large enterprises encounter.

“The government is struggling with digital transformation,” Issa said. “They’re stuck on all these legacy systems, and they’re not able to innovate as fast as the adversaries. So you’re seeing the likes of Iran and Syria and China and Russia and other Eastern Bloc countries start to build these offensive cyber capabilities. All you need is an internet connection, a bunch of skilled, dedicated resources, and now an entire country’s military cyber capability can rapidly grow. We don’t want that to outpace the United States.”

Part of the obstacle here is systems that run across tangled legacy and cloud infrastructure and mix structured and unstructured data and a wide range of coding languages. While there have been big gains in terms of protecting the underlying storage, Issa said attackers have increasingly focused on vulnerabilities in the applications.

“Encryption is something that everybody knows they need to do, but applying it without tripping over yourself is hard to do,” Issa said. “They turned to us because they’ve got all these disparate data types and they have all these unique types of storage. The problem is how to apply a uniform encryption strategy across all those diverse datasets.”

Issa said the emergence of the API economy has made such solutions far more accepted among big enterprises. They see APIs in general as a faster, more efficient way to build in functionality. Issa said applying that philosophy to security seemed like a natural evolution that not only eases the task but improves overall security.

“One of the other traditional challenges with encryption is when you deploy it somewhere and it breaks something,” he said. “And then you can’t deploy it in some sectors because the system is old. So you just apply it in two areas and then realize you’ve only applied encryption to 30% of your infrastructure. We enable a much more uniform approach.”

Ubiq got a boost earlier this month with a $ 6.4 million seed round. Okapi Venture Capital led the round, which included investment from TenOneTen Ventures, Cove Fund, DLA Piper Venture, Volta Global, and Alexandria Venture Investments. Ubiq plans to use the money for product development, building relationships with developers, and marketing.

“Our core focus is going to be on growing the platform, getting customer input, and making sure that we’re making the changes that our customers are asking for so we can run a very resilient, useful platform,” he said.

Let’s block ads! (Why?)

Big Data – VentureBeat

Alphabet’s Project Amber uses AI to try to diagnose depression from brain waves

November 3, 2020 Big Data

X, Alphabet’s experimental R&D lab, today detailed Project Amber, a now-disbanded project which aimed to make brain waves as easy to interpret as blood glucose. The goal was to develop objective measurements of depression and anxiety that could be used to support diagnoses, treatment, and therapies.

An estimated 17.3 million adults in the U.S. have had at least one major depressive episode, according to the U.S. National Institutes of Health. Moreover, the percentage of adults in the U.S. experiencing serious thoughts of suicide increased 0.15% from 2016-2017 to 2017-2018 — 460,000 more people than last year’s dataset. But with 1,000 possible symptom combinations, depression manifests differently in different people. Today’s assessments mostly rely on conversations with clinicians or surveys like the PHQ-9 or GAD-7.

The Amber team sought to marry machine learning techniques with electroencephalography (EEG) to measure telling electrical activity in the brain. Inspiration arose from the observation that game-like tasks can be used to gauge processing within the brain’s reward system. Brain response following a win in a game is subdued in people who are depressed, compared with those who are not.

X isn’t the first to apply machine learning algorithms to EEG readings. In a paper published last April, IBM researchers claimed to have developed an algorithm that could classify seizures with upwards of 98.4% accuracy. Indeed, EEGs have been widely used to study swallowing, classify mental states, and diagnose neuropsychiatric disorders such as neurogenic pain and epilepsy, as well as to classify emotions.

It took three years for the Amber team to create a low-cost, portable, research-grade system designed to make it easier to collect EEG data. The headset slips on like a swim cap and takes around three minutes to configure, using three sensors along the midline at Fz, Cz, and Pz (key channels, or electrodes, for assessments of reward and cognitive functions). It features an accompanying bioamp that can support up to 32 channels, which can be used to collect resting state EEG and event-related potentials with software that time-locks a task to the EEG measurement.

Beyond the headset, the Amber team explored how new approaches in machine learning could be used to reduce unwanted noise in EEG recordings. Collaborating with Alphabet’s deep learning research lab DeepMind, they adapted methods from unsupervised representation learning, demonstrating that approaches like autoencoders could be tapped to denoise EEG signals without a human in the loop. (Autoencoders learn representations for sets of data by ignoring noise.) In addition, the Amber team offered a proof of concept that it’s possible to extract features relevant to mental health that could be used to predict clinical labels like major depressive disorder and generalized anxiety disorder based on an interview by a mental health expert. Unlike previous studies, the Amber team claimed they were able to do this for an individual participant rather than a group.

“The methods were capable of recovering usable signal representations from single EEG trials,” X head Obi Felten explained in a blog post. “This means that it may be possible to derive clinically useful information from brain electrophysiology with far fewer data samples than what is traditionally used in research labs, which often rely on hundreds of experimental trials.”

The Amber team was ultimately unsuccessful in finding a single biomarker for depression and anxiety. However, despite their setbacks, they’ve released the hardware designs, visualizer, and stimulus tools they developed in open source on GitHub. As of this morning, the headset and software is available with the results of a study conducted with Florida State University. In addition, the Amber team is making a pledge not to assert its patents on Amber’s hardware and donating 50 unused EEG headsets to Sapien Labs, which runs the Human Brain Diversity Project supporting EEG research in low-income countries and with underrepresented groups.

“We hope that open-sourcing our EEG system and publishing our machine learning techniques will be of value not just to EEG experts, but also to the wider mental health research community who were perhaps put off by the complexity and cost of working with EEG before,” Felten wrote. “There are many pitfalls on the path to making tech-enabled mental health measurement work in the real world, and more research needs to be done … Addressing today’s challenges will require new partnerships between scientists, clinicians, technologists, policymakers, and individuals with lived experience.”

How startups are scaling communication:

The pandemic is making startups take a close look at ramping up their communication solutions. Learn how

Let’s block ads! (Why?)

Big Data – VentureBeat

Salesforce’s Simulation Cards spell out uses, risks, and bias to make AI models more transparent

October 20, 2020 Big Data

Salesforce recently open-sourced Foundation (formerly AI Economist), an AI research simulation for exploring tax policies. To accompany its release, the company this week published what it’s calling a Simulation Card, a file to document the use, risks, and sources of bias in published versions of the simulation.

Simulation Cards join ongoing efforts to bring transparency to historically black-box systems. Over the past year, Google launched Model Cards, which sprang from a Google AI whitepaper published in October 2018. Model Cards specify model architectures and provide insight into factors that help ensure optimal performance for given use cases. The idea of Model Cards emerged following Microsoft’s work on “datasheets for data sets,” or datasheets intended to foster trust and accountability through documenting datasets’ creation, composition, intended uses, maintenance, and other properties. Two years ago, IBM proposed its own form of model documentation in voluntary factsheets called “Supplier’s Declaration of Conformity” (DoC) to be completed and published by companies developing and providing AI.

The objective of Simulation Cards is similar to that of Model Cards and Data Sheets. However, Simulation Cards reflect the fact that simulations differ from trained models and datasets because they’re designed to create scenarios of interest, according to Salesforce. These scenarios can contain bias, which might be purposefully built-in or an unexpected side effect of the design choices made during creation. Because simulations create many datasets of various shapes and sizes, the potential for misuse is greater than that of a single fixed dataset that might contain bias.

Salesforce’s Simulation Cards spell out uses, risks, and bias to make AI models more transparent

Above: The Simulation Card for Foundation.

Image Credit: Salesforce

The Simulation Card for Foundation is divided into several sections: Simulation Details, Basic Information, Intended Use, Factors, Metrics, Quantitative Analyses, Ethical Considerations, and Caveats and Recommendations. Simulation Details provides the date of the simulation and the name of the publishing organization, as well as any keywords, licenses, contact information, and relevant version numbers. The Basic Information and Intended Use sections cover top-level info about the simulation and the applications for it that the coauthors had in mind. Factors canvases the modeling assumptions the simulations make, while Metrics and Quantitative Analyses outline the metrics used to measure the results. Finally, Ethical Considerations and Caveats and Recommendations provide guidelines for (or warnings against) applying the outputs to real-world systems.

It remains to be seen what sort of third-party adoption Simulation Cards might gain, if any, but Salesforce itself says it’s committed to releasing cards alongside future simulations. “We encourage researchers and developers to publish similar Simulation Cards for software releases, to broadly promote transparency and the ethical use of simulation frameworks. AI simulations offer researchers the power to generate data and evaluate outcomes of virtual economies that capture a part of the real world,” Salesforce wrote in a blog post. “An unethical simulation poses an order-of-magnitude larger ethical risk. As a result, our commitment to transparency is all that much more critical.”

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

IBM uses AI to evaluate risk of developing genetic diseases

August 20, 2020 Big Data

In a study published in the journal Nature Communications, scientists at IBM, the Broad Institute of MIT and Harvard, and health tech company Color find evidence that the presence of genetic mutations isn’t a reliable precursor to genetic diseases. They claim diseases can be so greatly influenced by other factors that the risk in carriers is sometimes as low as in that of noncarriers.

The research — which stems from a larger, three-year collaboration between IBM Research and the Broad Institute that was announced in 2019 — aims to support clinicians leveraging data to better identify patients at serious risk for conditions like cardiovascular disease. Insights could be useful in making health care and prevention decisions, helping clinicians choose whether to recommend imaging or more drastic, surgical interventions, like mastectomies.

In the course of the study, an IBM-led team developed models that analyze a person’s genetic risk factors, clinical health records, and biomarker data to more accurately predict the onset of conditions like heart attacks, sudden cardiac death, and atrial fibrillation. With Color, the researchers investigated whether polygenic background — the variants and factors within an individual’s genome — could influence the occurrence of disease in genomic conditions such as familial hypercholesterolemia, hereditary breast and ovarian cancer, and Lynch syndrome.

The coauthors analyzed de-identified records from over 80,000 patients across two large data sets, the UK Biobank and Color’s own. Among carriers of a monogenic risk variant, they identified “substantial variations” in risk based on polygenic background, implying carriers don’t always develop diseases. For example, the probability of developing coronary artery disease (CAD) by age 75 ranged from 17.5% to 77.9% for carriers of familial hypercholesterolemia. For comparison, CAD prevalence among noncarriers ranged from 13% if they had a low-risk polygenic score to 41% with a high-risk score.

“By leveraging large databases to combine and analyze medical and genomic data from tens of thousands of people, we have been able to shed significant new light on a number of serious, chronic diseases,” IBM Research principal scientist and coauthor Kenney Ng said. “Ultimately, our findings unveil a silver lining: Even if an individual carries a genetic mutation associated with one of these diseases, their absolute risk might not be as set in stone as previously thought. In fact, their absolute risk might be nearly equivalent to an individual who doesn’t carry the mutation at all — depending on other factors and mutations within their specific genome.”

IBM says future work will include investigating ways genomics, clinical data, and AI can be harnessed to develop new tools that offer health professionals insight into disease risk. The goal is to build algorithms that accurately indicate a predisposition to a health condition and make those tools available — including methods to calculate an individual’s risk of disease based on variants in a genome.

IBM previously collaborated with the Broad Institute in 2016. As part of a five-year project, the company sought to help researchers using AI and genomics better understand how cancers become resistant to therapies.

Let’s block ads! (Why?)

Big Data – VentureBeat

COVID-KG uses AI to scan thousands of studies to answer doctors’ coronavirus questions

July 3, 2020 Big Data

The number of studies about COVID-19 has risen steeply from the start of the pandemic, from around 20,000 in early March to over 30,000 as of late June. In an effort to help clinicians digest the vast amount of biomedical knowledge in the literature, researchers affiliated with Columbia, Brandeis, DARPA, UCLA, and UIUC developed a framework — COVID-KG, for “knowledge graph” — that draws on papers to answer natural language questions about drug purposing and more.

The sheer volume of COVID-19 research makes it difficult to sort the wheat from the chaff. Some false information has been promoted on social media and in publication venues like journals. And many results about the virus from different labs and sources are redundant, complementary, or even conflicting.

COVID-KG aims to solve the challenge by reading papers to build multimedia knowledge graphs consisting of nodes and edges. The nodes represent entities and concepts extracted from papers’ text and images, while the edges represent relations involving these entities.

COVID-KG ingests entity types including genes, diseases, chemicals, and organisms; relations like mechanisms, therapeutics, and increased expressions; and events such as gene expression, transcription, and localization. It also draws on entities annotated from an open source data set tailored for COVID-19 studies, which includes entity types like coronaviruses, viral proteins, evolution, materials, and immune response).

VB Transform 2020 Online – July 15-17. Join leading AI executives: Register for the free livestream.

COVID-KG extracts visual information from figure images (e.g., microscopic images, dosage response curves, and relational diagrams) to enrich the knowledge graph. After detecting and isolating figures from each document with text in its caption or referring context, it then applies computer vision to spot and separate non-overlapping regions and recognize the molecular structures within each figure.

COVID-KG provides semantic visualizations like tag clouds and heat maps that allow researchers to get a view of selected relations from hundreds or thousands of papers at a single glance. This, in turn, allows for the identification of relationships that would typically be missed by keyword searches or simple word cloud or heatmap displays.

In a case study, the researchers posed a series of 11 questions typically answered in a drug repurposing report to COVID-KG, like “Was the drug identified by manual or computation screen?” and “Has the drug shown evidence of systemic toxicity?” With three drugs suggested by DARPA biologists (benazepril, losartan, and amodiaquine) as targets, they used COVID-KG to construct a knowledge base from 25,534 peer-reviewed papers.

Given the question “What is the drug class and what is it currently approved to treat?” for benazepril, COVID-KG responded with:

The team reports that in the opinion of clinicians and medical school students who reviewed the results, COVID-KG’s answers were “informative, valid, and sound.” In the future, the coauthors plan to extend the system to automate the creation of new hypotheses by predicting new links. They also hope to produce a common semantic space for literature and apply it to improve COVID-KG’s cross-media knowledge grounding, inference, and transfer.

“With COVID-KG, researchers and clinicians are able to obtain trustworthy and non-trivial answers from scientific literature, and thus focus on more important hypothesis testing, and prioritize the analysis efforts for candidate exploration directions,” the coauthors wrote. “In our ongoing work we have created a new ontology that includes 77 entity subtypes and 58 event subtypes, and we are re-building an end-to-end joint neural … system following this new ontology.”

Let’s block ads! (Why?)

Big Data – VentureBeat

Dex-Net AR uses Apple’s ARKit to train robots to grasp objects

June 11, 2020 Big Data

UC Berkeley AI researchers are using an iPhone X and Apple’s ARKit to train a robotic arm how to grasp an object. It’s part of Dex-Net AR, a pipeline for using commodity smartphones for robotic grasping. ARKit creates point clouds from data generated by moving an RGB camera around an object for two minutes.

Robotic grasping is a particular robotics subfield focused on the challenge of teaching a robot to pick up, move, manipulate, or grasp an object. The Dexterity Network, or Dex-Net, research project at UC Berkeley’s Autolab dates back to 2017 and includes open source training data sets and pretrained models for robotic grasping in an ecommerce bin-picking scenario. The ability for robots to quickly learn how to grasp objects has a big impact on how automated warehouses like Amazon fulfillment centers can become.

In early experiments with eight objects in a laboratory, Dex-Net AR converted ARKit scans to depth maps for an ABB YuMi robot to grasp objects with a success rate of 95%. Each scan creates a point cloud.

“As the camera moves through space, the density of the point cloud increases, better detecting and defining the object’s surfaces for grasping,” a recently published paper detailing Dex-Net AR reads. “Dex-Net AR can generate grasps with accuracy similar to state-of-the-art systems that rely on expensive, industry grade depth sensors. Compared to depth camera systems that capture images from a fixed view, usually top-down, Dex-Net AR allows the user to move the smartphone camera all around the object, collecting three-dimensional point cloud data.”

VB Transform 2020 Online – July 15-17. Join leading AI executives: Register for the free livestream.

Dex-Net AR cleans up noise caused by estimation errors in ARKit point clouds using an outlier removal algorithm and k-nearest neighbor algorithm. The Dex-Net grasp planner then evaluates how the robot should pick up the object.

Since each ARKit scan took a fixed two minutes per object, in future efforts researchers will look for ways to scan objects more quickly. “[O]ne potential improvement is that we can try to bring down the amount of time in video capturing using a learning-based method to augment and complete the point cloud data given that only limited data are available,” the paper reads. Researchers also plan to explore how to better utilize the iPhone X depth-sensing cameras to collect cleaner point cloud data.

Dex-Net AR was introduced last week at the International Conference on Robotics and Automation (ICRA). Other papers published at the conference include works that explore ideal ways to walk for lower body skeletons for humans and four-legged robots. A Stanford lab shared a multi-drone management system that utilizes public buses to reduce delivery costs and energy consumption. Google Brain, Intel AI Lab, and Autolab also introduced Motion2Vec, AI trained AI for robotic surgery using video observation.

Let’s block ads! (Why?)

Big Data – VentureBeat

Google’s Objectron uses AI to track 3D objects in 2D video

March 12, 2020 Big Data

Coinciding with the kickoff of the 2020 TensorFlow Developer Summit, Google today published a pipeline — Objectron — that spots objects in 2D images and estimates their poses and sizes through an AI model. The company says it has implications for robotics, self-driving vehicles, image retrieval, and augmented reality — for instance, it could help a factory floor robot avoid obstacles in real time.

Tracking 3D objects is a tricky prospect, particularly when dealing with limited compute resources (like a smartphone system-on-chip). And it becomes tougher when the only imagery (usually video) available is 2D due to a lack of data and a diversity of appearances and shapes of objects.

The Google team behind Objectron, then, developed a toolset that allowed annotators to label 3D bounding boxes (i.e., rectangular borders) for objects using a split-screen view to display 2D video frames. 3D bounding boxes were overlaid atop it alongside point clouds, camera positions, and detected planes. Annotators drew 3D bounding boxes in the 3D view and verified their locations by reviewing the projections in 2D video frames, and for static objects, they only had to annotate the target object in a single frame. The tool propagated the object’s location to all frames using ground truth camera pose information from AR session data.

To supplement the real-world data in order to boost the accuracy of the AI model’s predictions, the team developed an engine that placed virtual objects into scenes containing AR session data. This allowed for the use of camera poses, detected planar surfaces, and estimated lighting to generate physically probable placements with lighting that matches the scene, which resulted in high-quality synthetic data with rendered objects that respected the scene geometry and fit seamlessly into real backgrounds. In validation tests, accuracy increased by about 10% with the synthetic data.

Better still, the team says the current version of the Objectron model is lightweight enough to run in real time on flagship mobile devices. With the Adreno 650 mobile graphics chip found in phones like the LG V60 ThinQ, Samsung Galaxy S20+, and Sony Xperia 1 II, it’s able to process around 26 frames per second.

The Objectron is available in MediaPipe, a framework for building cross-platform AI pipelines consisting of fast inference and media processing (like video decoding). Models trained to recognize shoes and chairs are available, as well as an end-to-end demo app.

The team says that in the future, it plans to share additional solutions with the research and development community to stimulate new use cases, applications, and research efforts. Additionally, it intends to scale the Objectron model to more categories of objects and further improve its on-device performance.

Let’s block ads! (Why?)

Big Data – VentureBeat

GitHub now uses AI to recommend open issues in project repositories

January 22, 2020 Big Data

Large open source projects on GitHub have intimidatingly long lists of problems that require addressing. To make it easier to spot the most pressing, GitHub recently introduced the “good first issues” feature, which matches contributors with issues that are likely to fit their interests. The initial version, which launched in May 2019, surfaced recommendations based on labels applied to issues by project maintainers. But an updated release shipped last month incorporates an AI algorithm that GitHub claims surfaces issues in about 70% of repositories recommended to users.

GitHub notes that it’s the first deep-learning-enabled product to launch on Github.com.

According to GitHub senior machine learning engineer Tiferet Gazit, GitHub last year conducted an analysis and manual curation to create a list of 300 label names used by popular open source repositories. (All were synonyms for either “good first issue” or “documentation,” like “beginner friendly,” “easy bug fix,” and “low-hanging-fruit.”) But relying on these meant that only about 40% of the recommended repositories had issues that could be surfaced. Plus, it left project maintainers with the burden of triaging and labeling issues themselves.

The new AI recommender system is largely automatic, by contrast. But building it required crafting an annotated training set of hundreds of thousands of samples.

GitHub began with issues that had any of the roughly 300 labels in the curated list, which it supplemented with a few sets of issues that were also likely to be beginner-friendly. (This included those that were closed by a user who had never previously contributed to the repository, as well as issues closed that touched only a few lines of code in a single file.) After detecting and removing near-duplicate issues, several training, validation, and test sets were separated across repositories to prevent data leakage from similar content, and GitHub trained the AI system using only preprocessed and denoised issue titles and bodies to ensure it detected good issues as soon as they’re opened.

In production, each issue for which the AI algorithm predicts a probability above the required threshold is slated for recommendation, with a confidence score equal to its predicted probability. Open issues from non-archived public repositories that have at least one of the labels from the curated label list are given a confidence score based on the relevance of their labels, with synonyms of “good first issue” awarded higher confidence than synonyms of “documentation.” At the repository level, all detected issues are ranked primarily based on their confidence score (though label-based detections are generally given higher confidence than ML-based detections), along with a penalty on issue age.

Data acquisition, training, and inference pipelines run daily, according to Gazit, using scheduled workflows to ensure the results remain “fresh” and “relevant.” In the future, GitHub intends to add better signals to its repository recommendations and a mechanism for maintainers and triagers to approve or remove AI-based recommendations in their repositories. And it plans to extend issue recommendations to offer personalized suggestions on next issues to tackle for anyone who has already made contributions to a project.

Let’s block ads! (Why?)

Big Data – VentureBeat

Ubisoft uses AI to teach a car to drive itself in a racing game

December 29, 2019 Big Data

Reinforcement learning, an AI training technique that employs rewards to drive software policies toward goals, has been applied successfully to domains from industrial robotics to drug discovery. But while firms including OpenAI and Alphabet’s DeepMind have investigated its efficacy in video games like Dota 2, Quake III Arena, and StarCraft 2, few to date have studied its use under constraints like those encountered in the game industry.

That’s presumably why Ubisoft La Forge, game developer Ubisoft’s eponymous prototyping space, proposed in a recent paper an algorithm that’s able to handle discrete, continuous video game actions in a “principled” and predictable way. They set it loose on a “commercial game” (likely The Crew or The Crew 2, though neither is explicitly mentioned) and report that it’s competitive with state-of-the-art benchmark tasks.

“Reinforcement Learning applications in video games have recently seen massive advances coming from the research community, with agents trained to play Atari games from pixels or to be competitive with the best players in the world in complicated imperfect information games,” wrote the coauthors of a paper describing the work. “These systems have comparatively seen little use within the video game industry, and we believe lack of accessibility to be a major reason behind this. Indeed, really impressive results … are produced by large research groups with computational resources well beyond what is typically available within video game studios.”

The Ubisoft team, then, sought to devise a reinforcement learning approach that’d address common challenges in video game development. They note that data sample collection tends to be a lot slower generally, and that there exist time budget constraints over the runtime performance of agents.

Their solution is based on the Soft Actor-Critic architecture proposed early last year by researchers at the University of California, Berkeley, which is more sample-efficient than traditional reinforcement learning algorithms and which robustly learns to generalize to conditions that it hasn’t seen before. They extend it to a hybrid setting with both continuous and discrete actions, a situation often encountered in video games (e.g., when a player has the freedom to perform actions like moving and jumping, each of which are associated with parameters like target coordinates and direction).

The Ubisoft researchers evaluated their algorithm on three environments designed to benchmark reinforcement learning systems, including a simple platformer-like game and two soccer-based games. They claim that its performance fell slightly short of industry-leading techniques, which they attribute to an architectural quirk. But they say that in a separate test, they successfully used it to train a video game vehicle with two continuous actions (acceleration and steering) and one binary discrete action (hand brake), the objective being to follow a given path as quickly as possible in environments the agent didn’t encounter during training.

“We showed that Hybrid SAC can be successfully applied to train a car on a high-speed driving task in a commercial video game,” wrote the researchers, who futher noted that their approach can accommodate a wide range of potential ways for an agent to interact with a video game environment, such as when the agent has the same inputs as a player (whose controller might be equipped with an analog stick that provides continuous values and buttons that can be pressed to yield discrete actions through combinations). “[This demonstrates] the practical usefulness of such an algorithm for the video game industry.”

Let’s block ads! (Why?)

Big Data – VentureBeat

Tag Archives: Uses

Insightin Health raises $12 million for AI that uses big data to guide patients’ decisions

VentureBeat

How Ubiq Security uses APIs to simplify data protection

Lessons from the government

Alphabet’s Project Amber uses AI to try to diagnose depression from brain waves

Salesforce’s Simulation Cards spell out uses, risks, and bias to make AI models more transparent

IBM uses AI to evaluate risk of developing genetic diseases

COVID-KG uses AI to scan thousands of studies to answer doctors’ coronavirus questions

Dex-Net AR uses Apple’s ARKit to train robots to grasp objects

Google’s Objectron uses AI to track 3D objects in 2D video

Ubisoft uses AI to teach a car to drive itself in a racing game

Recent Posts

Categories

Archives