Tech giant Microsoft acquires artificial intelligence startup Suplari that specializes in developing solutions for corporate expenditure and cash flow analysis.
Suplari’s platform is also capable of predicting fund inflow and trend tracking. With this acquisition, Microsoft plans to integrate Suplari’s artificial intelligence-powered platform with its cloud-based Dynamic 365 solution.
The merger would help businesses to maximize financial visibility by analyzing current data and previous trend patterns from various data sources using artificial intelligence technology.
According to Microsoft, more than half of the businesses worldwide would use artificial intelligence cloud services for financial analysis needs by the year 2022.
Vice President of Dynamic 365 at Microsoft, Frank Weigel, said, “Together with Dynamics 365, the Suplari Spend Intelligence Cloud will help customers maximize financial visibility by using AI to automate the analysis of current data and historical patterns from multiple data sources. This acquisition will further empower Microsoft to help our customers turn data into actionable insight.”
He further added that this advancement would help enterprises to move beyond transactional financial management to artificial intelligence solutions that would enable them to make informed decisions, scrutinize risks, and reduce supplier costs.
Sulpari is a Seattle-based startup founded in the year 2016 by Nikesh Parekh that specializes in developing artificial intelligence solutions to allow businesses to make faster purchasing decisions using antiquated enterprise systems.
CEO of Sulpari, Nikesh Parekh, said, “We are excited for the new road ahead with Microsoft. I am ecstatic to report that 100pc of our team is continuing to work together to extend Suplari’s Spend Intelligence Cloud at Microsoft.”
He also mentioned that now the consumers can expect the company to develop better predictive and perspective insights generating platforms for the financial needs of various businesses.
Covariant, a leading AI Robotics company, raised $40 million in Series B funding in May 2020. Just after a year, the Berkeley-based AI start-up, Covariant raises $80M Series C funding, bringing its total capitalization to $147 million within two years of the company coming out from stealth mode.
The Series C round was led by returning investor, Index Ventures, with the additional participation of Radical Ventures and Amplify Partners. Covariant also added Temasek and Canada Pension Plan Investment Board (CPP Investments) as new global investors with this latest round of funding.
“With Covariant rolling out multiple applications in warehouses across Europe, North America, and Asia-Pacific over the last year, it’s the first time that AI Robotics has been successful at this scale with such variability. Covariant consistently outperforms the competition in tests by prospective clients to benchmark autonomy in real-world operations,” said Mike Volpi of Index Ventures.
Covariant was founded in 2017 by OpenAI along with researchers from the University of California, Berkeley. Last year, they deployed Covariant Brain, which was described as “universal AI that enables robots to see, reason, and act autonomously in the real world.” It’s currently active in markets across Europe, Asia, and North America. Covariant has deployed its AI and robotics technology across various sectors from groceries, Industrial supply, pharmaceuticals, parcels, general merchandise, health, beauty, and fashion.
The company currently has under 80 employees, and part of the funding will grow the team globally. Covariant has added high-profile team players to its management for meeting challenges: Ally Lynch as head of Marketing, Raghavendra Prabhu as head of Engineering and Research, and Sam Cauthen as head of People. Meanwhile, parts of the $80 million funding will continue to be invested in AI Robotics research and development (R&D) and accelerate bringing AI into the physical world.
Microchip manufacturing giant Intel launches its new free initiative ‘AI for All’ in collaboration with the Central Board of Secondary Education (CBSE) and the Ministry of Education to teach individuals the basics of artificial intelligence in India.
It is a 4 hours self-paced program that anyone can attend for free. With this new initiative, Intel plans to educate more than 1 million individuals in a span of one year.
Director of APJ and Global Partnerships and Initiatives at Intel, Shweta Khurana, said, “Artificial Intelligence has the power to drive faster economic growth, address population-scale challenges and benefit the lives and livelihoods of people. The AI For All initiative based on Intel’s AI For Citizens program aims to make India AI-ready by building awareness and appreciation of AI among everyone.”
She further mentioned that this effort of Intel shows its willingness to work with the government of India towards making the country digitally empowered and unlock the full potential of artificial intelligence in India.
The free course is divided into two sections that would talk about artificial intelligence awareness and its applications, respectively. It would also focus on resolving common misconceptions people have regarding artificial intelligence.
The COVID-19 pandemic has forced businesses to adopt a digital approach, which has accelerated the growth of artificial intelligence technologies in every industry. This course will help the citizens of India to understand the new normal in a better way and adapt accordingly.
Interested candidates can enroll themselves through the official website of CBSE.
In a recent event, the Prime Minister of India announced the launch of the AI for All initiative along with another program named ‘SAFAL.’ CBSE, in a statement, mentioned, “SAFAL will be conducted on a pilot basis in CBSE schools for students in Grades 3, 5, and 8 during the academic year 2021-22, in key curricular areas of Language, Mathematics, and EVS/Science.”
On Tuesday, DeepMind released a preprint on “Open-Ended Learning Leads to Generally Capable Agents,” outlining its initial efforts toward training an agent capable of playing a variety of games without relying on human interaction data. The team built a large training environment called XLand, that produces multiplayer mini-games within stable, “human-relatable” digital 3D scenarios on its own. This environment enables the development of new learning algorithms that dynamically control how an agent learns and the games on which it trains, allowing for the mass training of AI agents to do tasks of variable complexity.
The objective behind creating XLand is to overcome the limitations of training artificial intelligence models (and robots) using reinforcement learning. The reinforcement learning algorithms learn how to do things by synthesizing input from a huge dataset, identifying patterns, and using those patterns to generate educated guesses about fresh data. Simply put, the algorithm learns to make a series of decisions depending on the feedback it receives from its computational environment. A reinforcement learning model is rewarded for generating good predictions and punished for producing bad ones. Over time, it arrives at the optimal solution by attempting to maximize the cumulative reward.
However, the problem with this form of training is that the models are generally trained in a limited set of scenarios. Hence, if the same models are presented with a slightly different set of environments they may struggle to adapt to these environments nor produce satisfactory outcomes.
Therefore, rather than training agents to do a narrow set of activities, the DeepMind research team has discovered a universe of scenarios that may be produced procedurally. Each AI player’s aim is to maximize prizes, and each game determines the players’ unique awards. Deepmind also used population-based training (PBT) to prevent training dead ends. Population-based training is a neural network training approach that allows an experimenter to quickly select the optimum collection of hyperparameters and models for the job.
The AI agents in DeepMind’s XLand operate as a basic body in continuously changing digital surroundings that the agent views from a first-person perspective. Simple problems such as “Stand next to the purple cube,” “Bring the yellow cube onto the white corridor,” and linked conditions like “Stand next to the purple cube or in the red hallway,” were among the game tasks generated procedurally in the XLand. These tasks enable the AI agents to train themselves and generate experience by performing them.
The agents sense their surroundings by observing RGB images and receive a text description of their goal, direct feedback on success or failure follows at defined time intervals. There are other AI agents with similar or opposing aims in many of the generated games. Also, an AI agent may interact with interactive elements such as spheres, cubes, and ramps by employing tools that allow it to pick up or freeze them.
According to DeepMind, after five generations, AI agents exhibit continual breakthroughs in learning and skills that were previously unknown. In those five iterations, each AI agent has completed over 200 billion training steps due to 3.4 million distinct assignments and has played nearly 700,000 games in 4,000 different XLand environments.
The AI agents demonstrated general behavior tendencies such as exploration, like changing the state of the environment until they obtained a rewarding condition, after only 30 minutes of intense training on a new task. DeepMind reported that these agents were aware of the fundamentals of their bodies, the passage of time, and the high-level structure of the games they were playing.
Last month DeepMind claimed that reinforcement learning was enough to achieve General AI. Now it also admits that the above feat would not have been possible using the reinforcement learning method alone – thus paving way for zero-shot learning. In its blog, it wrote, “Instead of learning one game at a time, these [systems] would be able to react to completely new conditions and play a whole universe of games and tasks, including ones never seen before.”
Researchers from Skoltech Institute of Science and Technology, Lomonosov Moscow State University, and the Syntelly startup have created and trained a neural network that generates names for organic compounds using the IUPAC nomenclature system.
In their research published on Nature under Scientific Report, the team mentions creating a Transformer-based artificial neural network approach for translating between SMILES and IUPAC chemical notations: Struct2IUPAC and IUPAC2Struct. The Struct2IUPAC converts SMILES strings to IUPAC names and IUPAC2Struct performs the reverse conversion.
IUPAC or International Union of Pure and Applied Chemistry was founded in 1919 to harmonize the chemical naming of elements and organic compounds. For instance, in the IUPAC terms, sucrose is called (2R,3R,4S,5S,6R)-2-[(2S,3S,4S, 5R)-3,4-dihydroxy-2,5-bis(hydroxymethyl)oxolan-2-yl]oxy-6-(hydroxymethyl)oxane-3,4,5-triol, and paracetamol, the active ingredient of antipyretic drugs like Tylenol, is N-(4-hydroxyphenyl)acetamide. Since the IUPAC name comprises representing organic molecules’ names in the form of chemical structures using numbers and long names, it seems inconvenient to remember. Omitting even a single digit or symbol is unacceptable in the scientific domain.
Hence we have SMILES, or Simplified Molecular Input Line Entry System, which was created to make chemical information processing easier for both humans and computers. For example, Ethanol is written as CCO, which represents the molecule’s fundamental backbone, without any hydrogens: i.e., a carbon bonded with a carbon bonded to an oxygen. The best advantage to SMILES nomenclature is that many SMILES strings can describe the same molecule. For Ethanol, OCC and C(O)C are both acceptable.
According to Skoltech research scientist Sergey Sosnin, the team initially wanted to create an IUPAC name generator for Syntelly. However, they soon realized that it would take the team more than a year to create an algorithm by digitizing the IUPAC rules. Therefore they decided to leverage their knowledge and expertise in neural network solutions instead. Sergey is also the lead author of the study and co-founder of the Syntelly startup.
The team used the standard Transformer architecture with six encoder and decoder layers and eight attention heads as the basis for their research. The encoder layer creates an encoded representation of the words in the input data (latent vector or context vector). When a latent vector is provided to the decoder, it creates a target sequence by predicting the most likely word for each time step that pairs with the input word. Also, the Transformer uses an attention mechanism that looks at an input sequence and decides at each step which other parts of the sequence are important. This helps the neural network models to selectively focus on certain parts of their input and thus reason more effectively.
The team trained the Struct2IUPAC to convert a molecule’s structural representation to an IUPAC name and IUPAC2Struct for vice versa. They used PubChem, the world’s biggest open chemical library with over 100 million organic chemicals, to serve as the basis for the new network’s training and testing. The transformer neural network Struct2IUPAC learned to convert the names with almost 98.9% accuracy (1075 mistakes per 100,000 molecules) on a subset of 100,000 random organic molecules from the test set within six weeks of designing.
In recent years, the use of neural network approaches for solving chemical issues has grown in popularity. By treating molecules and reactions like words and sentences, they have found ways to get the computer to understand the chemical compounds. Yet, despite its enormous scope, the technology is still in its infancy.
Sergey says, “We have shown that neural networks can cope with exact problems, disproving the formerly prevalent notion that they should not be used for this kind of problem. Replacing a word with a synonym is possible in machine translation, whereas a single wrong symbol results in an incorrect molecule in our task. Yet, Transformer successfully copes with this task.”
The new solution has been integrated into the Syntelly platform and is available online. The researchers anticipate that their technique will be useful for converting between chemical notations as well as other technical notation-related activities like formula synthesis and software translation.
OpenAI released Triton 1.0, an open-source Python-like programming language for GPUs. It serves researchers with no CUDA (Compute Unified Device Architecture) experience to write highly efficient GPU codes. Triton enables peak hardware performance with relatively little effort and is 2x more efficient than equivalent Torch implementations.
In recent times, Deep Neural Networks (DNN) models have outperformed across many domains ranging from natural language processing to computer vision. However, deep learning neural models undergo high computation with parallel work, thereby requiring multi and many-core processors. Such High-Performance Computing (HPC) needs have increased the demands for GPUs to compute the processing of large data at a rapid speed.
Modern research in deep learning is implemented using a combination of the native framework of operators, which may require the creation of many temporary tensors. This approach lacks flexibility, is too verbose, and degrades the performance of the neural network. However, OpenAI’s Triton mitigated this issue by providing an intermediate language and compiler.
Triton’s success lies in the modular system architecture that is centered around Triton-IR, allowing their compiler to automatically perform a wide variety of important program optimizations. They had to revisit the traditional “Single Program, Multiple Data” (SPMD) thread execution model for GPU, and proposed a block algorithm, useful while performing sparse operations. This Block-based algorithm aggressively optimizes programs for data locality and parallelism.
OpenAI’s Triton aims to provide an open-source environment to write code faster with higher productivity and flexibility than CUDA and other existing DSLs (Domain Specific Language) respectively. Currently, Triton is compatible only with Linux and supports NVIDIA GPUs (Compute Capability 7.0+) hardware. The next release may support AMD GPUs, CPUs, and the foundations of this project are described in the following MAPL2019 publication: Triton: An Intermediate Language and Compiler for Tiled Neural Network Computations.
Defense technology startup Shield AI acquires Martin UAV through a definitive agreement on 28th July 2021. Martin UAV is a Texas-based startup that specializes in developing unmanned aircraft and provides cost-effective aerial services to its customers.
Shield AI aims to integrate Martin UAV’s platform with its own solution named Hivemind to develop airplanes capable of performing vertical landing and takeoffs for the United States Army and airforce.
Shield AI’s platform uses reinforcement learning, computer vision algorithms, and artificial intelligence to train various sorts of crewless vehicles to perform military operations. Martin UAV has developed one of the world’s best unmanned aerial vehicles that are capable of delivering a whopping eleven hours of fly time and is equipped with a single ducted thrust vectoring fan.
Company officials claim that its aerial vehicle can fly more than ten times than its competitors. The vehicle can carry a maximum of twenty-five pounds of payload.
Chief Executive Officer of Martin UAV, Ruben Martin, said, “GPS and communications on the battlefield are no longer assured. A great aircraft without an AI to make intelligent decisions will be sidelined against China, Russia, and an increasing number of adversaries who are fielding electronic warfare and anti-air systems. Shield AI is one of the only companies that has operationalized advanced aircraft autonomy on the battlefield.”
He further mentioned that this acquisition would make the company’s V-BAT UAV the world’s first vehicle capable of performing sustained operations in denied environmental conditions. Martin UAV’s vehicles have already been tested for two years by the US military to perform numerous operations. V-BAT also won a competitive selection round of the US military earlier this year.
Co-founder of Shield AI, Brandon Tseng, said, “Expeditionary. Intelligent. Collaborative. Expeditionary means capability on edge, within the control of the units who need it most. V-BAT is expeditionary today. Intelligent means aircraft that make their own decisions to execute the commander’s intent to accomplish missions with or without reach-back.”
Researchers at McGill University are now using new artificial intelligence technology to predict suicidal behavior in students. The COVID-19 pandemic has severely affected the mental health of students that has increased their suicidal behavior.
A research team has been assembled from various Universities to develop artificial intelligence algorithms to recognize factors that would help predict early signs of suicidal behavior. This technology will allow doctors to treat such students at an early stage to improve their mental health.
Massimiliano Orri of McGill University said, “Many known factors can contribute to the increased risk in university students, such as the transition from high school to college, psychosocial stress, academic pressures, and adapting to a new environment. These are risks that have also been exacerbated by the health crisis triggered by the COVID-19 pandemic, although there is no clear evidence of an increase in suicide rates during the pandemic.”
According to a Ph.D. student of the University of Bordeaux, suicide is the world’s second most significant cause of death among 15 to 24-year-old individuals. Researchers are developing the artificial intelligence algorithm after analyzing the data of more than five thousand university students in France from 2013 to 2019.
The researchers were able to detect four factors like anxiety, self-esteem, and depressive symptoms that are capable of recognizing suicidal behavior with an accuracy of 80%. According to a survey conducted by the researchers, around 17% of students, both men, and women had shown suicidal behavior while completing their degree courses.
A professor at the University of Bordeaux, Christophe Tzourio, said, “This research opens up the possibility of large-scale screening by identifying students at risk of suicide using short, simple questionnaires, in order to refer them to appropriate care.”
He further added that this technology could provide an alternative to mental health assessments by doctors for students.
Researchers at Harvard University are now using artificial intelligence to search alien technologies in space. Under the Galileo Project of Harvard University, scientists are trying to locate extinct and active extraterrestrial civilizations in deeper parts of the universe.
The research is using a number of high-power telescopes as its primary instrument to gather evidence of the presence of aliens. Scientists are using artificial intelligence algorithms to identify alien-built satellites and unidentified aerial phenomena (UAP).
The research has already received funding of $1.75 million. Prof. Abe Loeb, Head researcher of project Galileo, said, “We can no longer ignore the possibility that technological civilizations predated us.”
According to the scientists, the data gathered from numerous telescopes will be scanned and analyzed by an artificial intelligence algorithm to identify alien existence. Loeb said, “Science should not reject potential extraterrestrial explanations because of social stigma or cultural preferences that are not conducive to the scientific method of unbiased, empirical inquiry.”
He further mentioned that now scientists must start looking through new telescopes, both literally and figuratively. The impact of the finding of any extraterrestrial establishment will be enormous as it will affect our current technological approach and our society.
Loeb had earlier proposed his theory that the comet Oumuamou that crossed the orbit of Earth in 2017 was an alien technology. He believes that such incidents will boost scientists’ willingness to conduct further research on alien existence.
The project Galileo also aims to study interstellar objects that enter our solar system and identify alien satellites that spy on the Earth. Abe Loeb has collaborated with renowned scientists like Stephen Hawkings and has published more than a hundred scientific research papers till date.
“Science should not reject potential extraterrestrial explanations because of social stigma or cultural preferences that are not conducive to the scientific method of unbiased, empirical inquiry,” said Loeb in a recent statement.
Microsoft partnered with OpenAI and developed GitHub Copilot, which uses GPT-3 algorithm to suggest users with code generation. Many GitHub users have enjoyed Copilot as a pair programmer, and the majority of them embrace the suggestions while coding. However, the SendGrid engineer reported the first bug showcasing the issue of leaking sensitive and functional API keys, thereby giving access to databases.
Sam Nguyen, a software engineer of SendGrid, got a list of secret API keys when he asked the AI tool for the same. API keys are simple encrypted strings useful for accessing databases. The developer opened a request reporting this concern with a screenshot showcasing at least four proposed keys. Github CEO Nat Friedman has acknowledged the issue stating that “these secrets are almost entirely fictional, synthesized from the training data.”
Although Github’s team is working on this issue, it has ignited many open source developers to migrate from Github. Developers envy GitHub Copilot AI and claim this tool uses copyrighted source code in an unauthorized and unlicensed way.
One of the developers said “This product injects source code derived from copyrighted sources into their customers’ software without informing credits of the licensed source code. This significantly violates the terms of copyright holder’s work.” Currently, Microsoft has released a public version of Github Copilot, which is trained from codes from the public repositories of GitHub.
It laters plans to release a commercial product version, supporting enterprises in understanding their programming styles. This AI technology will not only be limited for Microsoft as OpenAI CTO Greg Brockman said “they will be releasing Codex model this summer for third party developers to tailor their own application.”
