Image file photo (REUTERS/Kacper Pempel/Files)
The Pentagon’s science and technology research arm is launching a powerful push to a new level of sophisticated artificial intelligence, referred to the integration of advanced levels of “machine learning” introduce more “adaptive reasoning” and help computers determine that it is more subjective phenomena.
It is the 3rd Wave (Defense Advanced Research Projects Agency program to exploit the rapid advances in AI to help in the training data to the computer of the analysis more reliable for human operators, Director Steven Walker recently told a small group of reporters.
DARPA scientists explain the rapidly evolving 3rd wave efforts, such as improving the ability of AI-based technology to allow for much more sophisticated “contextual explanatory models.”
While the people will still be needed in many cases, the 3rd Wave can be described as the introduction of a new ability to not only provide answers and interpretations, but also use machine learning to reason in the context and explain the results,” DARPA Deputy Director Peter Highnam said.
In short, the 3rd Wave can explain the reason “why” they came to the conclusion reached, something that provides a breakthrough level of human-computer interface, he added.
“When we talk about the 3rd wave, we are focused on contextual reasoning, and adaptation. It requires less data to training,” Highnam said.
This not only makes provisions more reliable, but massively increases a possibility of more subjective interpretations by understanding how different words or data sets relate to each other.
A computer can only draw from the information is entered or given, by and large. While adding seemingly unlimited amounts of data almost instantaneously, AI-driven analysis can be faced with challenges as well as elements of the underlying stored data change for a specific reason. It is precisely this situation that the 3rd Wave is intended to address.
“If the underlying data changes then your system is not trained against that,” Highnam explained.
For example, 3rd wave adaptive reasoning will enable computer algorithms to distinguish the difference between the use of “principal” and “principle” on the basis of an ability to analyse the surrounding words and the determination of the context.
This level of analysis naturally creates a much higher level of reliability and nuance as it can be empowerment of people with a much deeper understanding of the detailed information that they can find.
“That is the future – building enough AI in the machines that they can actually communicate, share information and network with the machine speed in real-time,” Walker said.
Yet another example of the emerging advanced levels of AI would be an ability to organize hours of drone-collected video-very quickly – and the determination of the moments that is important for the human decision-makers. This exponentially increases the speed of human decision-making, a factor that can easily determine life or death in battle.
“In a warfare scenario, people are too trusting when the computer gives them an answer…by using contextual reasoning,” Highnam said.
Given these emerging 3rd Wave of advances, making more subjective decisions will increasingly be on a realistic element of AI-functional competence. For this reason and others, DARPA works closely with the private sector to strengthen cooperation with silicon valley and defense-industry partners as a way to identify and apply the latest innovations.
DARPA ‘ s of the 1st, 2nd and 3rd Wave of AI
The third wave, described in DARPA materials ” “contextual explanatory models’, and a much higher level of machine learning, it is intended to build on the 1st and 2nd Waves of DARPA’s previous AI progress.
The 1st Wave, depending on the available DARPA information “can reason about the narrowly-defined problems.” Although certain elements of the learning capability, it is described as having a “poor level of security.”
This points to the principle challenge of AI, namely, the promotion of a ability to generate “trust” or reliability that the process through which he discovered new patterns, gives you answers and compares new data against volumes of historical data are not correct. Given this challenge, especially in the existing models of AI-integration would have trouble adjusting to the change of data, or the determination of sufficient context.
The 2nd Wave “of statistical models and lead them on big data”, but has minimal reasoning, DARPA materials to explain. This means algorithms are able to recognize new information and often in a broader context, in connection with an existing database.
The 2nd Wave, therefore, can often determine the meaning of previously-recognized words and information is by examining the context and the execution of certain levels of interpretation. AI-enabled computer algorithms, in this phase, are able to accurately analyze the words and information by placing them in context with the surrounding data and concepts.
With this 2nd wave, however, DARPA scientist explain that there are restrictions with regard to the reliability of the interpretation, and an ability to react to new information in some cases; this can make the measurement less reliable. Highnam explained this as less of an opportunity to work out of existing data or as and when new information changes. Therefore, this Wave is described by DARPA information to have “a minimum of reasoning.”
Can AI Make Subjective Observations?
Raytheon, for example, is currently exploring a joint research deal with the Navy to explore predictions on the basis of the maintenance and the training of algorithms to perform real-time analysis on other complex problems. It is a 6-month Cooperative Research and Development Agreement (CRADA) to explore a comprehensive new AI-based applications, the company developers said.
Raytheon developers are naturally reluctant to give up any specific issues or platforms they are engaged with the Navy, but said they were looking for better AI so that large warfigthing systems, weapons and networks.
Todd Probert, Raytheon, Vice-President of the Mission Support and Modernization, told the Warrior Maven in an interview what their commitment is to work on initiatives that compliment DoD, current AI push.
“A part of the commitment of the AI is about gaining the confidence to trust that the AI as edit operations, and then break it down even further,” Probert said. “We’re training algorithms to do the work of man.”
It is interesting that the types of progress made possible by a 3rd Wave brings the prospect of engineering AI-driven algorithms to interpret subjective nuances. For example, things like certain philosophical concepts, emotions and psychological nuances influenced by past experiences, seems to be the kind of thing that computers would not be able to interpret.
Although this is, of course, still true in many ways, as even the most advanced algorithms still are not parallel human cognition, emotion, in some respects, the AI is more and more able to be on a more subjective terms, Probert said.
Probert explained that advanced AI is capable of certain types of intentions, emotions and prejudices by means of a ability to collect and organize information related to word selection, speech recognition, patterns of expression and intonation as a way to distinguish on a more subjective phenomena.
Also, if a system has a sufficiently large database, perhaps including the prior expressions, writing, or information with respect to the new information, placing of new words, phrases, and incoming data in a broader, more subjective context, Probert explained.
AI & counter – Terrorism- Torres AES
Other partners are the use of new levels of AI to strengthen counter-terrorism investigations and cyber forensics. For example, a U.S.-based global security company supporting the DoD, the U.S. State Dept. and friendly foreign governments, Torres Advanced enterprise Solutions, makes use of advanced levels of AI to discover otherwise obscured or hidden communication between terrorist groups, transnational criminals, or other AMERICAN opponents.
While many of the details of this type of AI application, the company’s developers say, of course, are not available for reasons of safety, Torres cyber forensics experts say advanced algorithms to find associations and “digital footprints” associated with bad actors or enemy activity with newer methods of AI.
As part of the cyber forensics training of the USA. and the US-nato counter-terrorism forces, Torres prepares cyber warriors and investigators to exploit the AI. Torres does cyber forensics training of the USA-nato-Argentine and the Paraguayan counter-terrorism officials, for example, often look to crack down on terrorist financial activity in the more loosely based on the “tri-border” region connecting Paraguay, Argentina and Brazil.
“The system that we train builds up in the AI, but does not eliminate the man. AI-compatible algorithms can identify direct and indirect digital relationships among bad actors,” said Jerry Torres Torres AES CEO.
For example, AI can use adaptive reasoning to distinguish the relationships between the locations, names, email addresses or bank accounts are to be used by bad actors.
To illustrate some of the effective applications of AI for this kind of efforts, Torres pointed to a own software called Maltego was used for the open-source intelligence analysis, and forensic research.
“AI can be a great asset that our defensive cyber systems to learn about the attackers by increasing the knowledge base of each attack, and the launch of intelligent counter-attacks to neutralize attackers, or feign an attack to the attacker to expose itself. AI is essential for the control of the attackers,” Torres added.
The software uses AI to search for relationships in a variety of online services to the use of social media, domains, groups, networks, and other areas of research relevance.
The Growing Impact of AI
AI has advanced rapidly to unprecedented levels of autonomy and machine learning, where algorithms are readily able to assimilate and analyze new patterns and new information, context and to compare with enormous amounts of data. Many now follow the seemingly countless applications of this whole military networks, data, systems, weapons, and large platforms.
Computer autonomy currently performs procedural functions, the organisation of information and makes for an incredible speed of processing is designed to have a much faster decision-making and the solving of problems by the person performing command and control. While AI can prove a seemingly infinite amount of great importance in short order – or almost immediately – the human cognition is still required in many cases to integrate less “hard” variables, solving dynamic problems, or even a more subjective analysis.
When it comes to the current and emerging platforms, there is already a lot of progress in the area of AI; the F-35s “sensor fusion” is based on the early stage of AI, Navy Ford-Class carriers using higher levels of automation for performing on-board functions, and Virginia-Class Block III attack submarines draw on the touch screen for fly-by-wire ‘ technology to provide more autonomy to underwater navigation.
Other cases are the Army of the current experiments with IBM’s AI-enabled computer, Watson, which among other things can be used to wirelessly perform real-time analyses on the control of relevant maintenance details on Stryker vehicles. In a manner somewhat analogous to that of the company named C3IOT makes use of AI-authorised real-time analyses of air-conditioned-based maintenance on Air Force F-16’s.
“Despite the higher levels of autonomy, at the end of the human will make the decision, use of computers as partners. We see the future much less to do with machines to do everything, but the people and the machines work together to fight the next battle,” Highnam explained.
Ultimately, Highnam said: —“war in the Future will be about speed – turning information into knowledge faster.” —
More Weapons and Technology – WARRIOR MAVEN (CLICK HERE