Looking at quantum computing-fueled applications of the future, we much of the time look to the innovation’s capability to take care of computationally-intensive mathematical problems, which could lead to breakthroughs in drug discovery, logistics, cryptography, and finance.

A research paper by Bernhard Dieber and different scholastics entitled Quantum Computation in Robotic Science and Applications, researches how quantum computing could augment numerous operations where robots are confronted with intensive computational assignments, where commonly broadly useful GPUs have been utilized to deal with intensive tasks.

While we may not see the appearance of quantum-fueled robots in the coming decade, the paper refers to how the rise of cloud-based quantum computing services and even quantum co-processors (QPUs) could work coupled with traditional CPUs to propel the improvement of much increasingly powerful and smart robots.

Australian physicists state they have adapted methods from autonomous vehicles and robotics to effectively evaluate the performance of quantum gadgets. A University of Sydney team reports that its new methodology has been indicated tentatively to outflank simplistic characterisation of these situations by a factor of three, with a lot higher outcome for increasingly complex simulated environments. Lead creator Riddhi Gupta says one of the hindrances to creating quantum computing systems to useful scale is beating the blemishes of hardware.

Qubits – the fundamental units of quantum technology are exceptionally delicate to disturbances from their environments, for example, electromagnetic noise and show performance varieties that lessen their usefulness.

To address this, Gupta and associates took strategies from old style estimation utilized in robotics and adapted them to improve hardware performance. This is accomplished through the proficient automation of procedures that map both environment of and performance variations across huge quantum gadgets.

Conventional AI, as opposed to current machine learning applications, depends on formal knowledge representations like rules, realities and algorithms so as to improve the robot behavior or copy intelligent behavior.

Artificial intelligence applications are as often as possible utilized in robotics technology, similar to path planning, the derivation of goal-oriented action plans, system diagnosis, the coordination of different specialists, or thinking and reasoning of new knowledge. A significant number of these applications use varieties of ignorant (visually impaired) or informed (heuristic) search algorithms, which depend on crossing trees or diagrams, where every node represents a potential state in the search space, associated with further follow-up states.

Quantum computing can fill in as an option for pretty much every search algorithm utilized in robotics and AI applications and decrease unpredictability. For graph search, for instance, there is a quantum alternative based on quantum random walks.

In robotics, Gupta says, machines depend on simultaneous localisation and mapping (SLAM) algorithms. Gadgets like automated vacuum cleaners are ceaselessly mapping their surroundings and then evaluating their area within that environment so as to move. The trouble with adjusting SLAM algorithms to quantum frameworks is that if you measure, or characterise, the performance of a solitary qubit, you obliterate its quantum data.

Gupta has built up a versatile algorithm that measures the performance of one qubit and utilities that data to assess the capacities of nearby qubits. “We have called this Noise Mapping for Quantum Architectures.,” she says. “Instead of gauging the old-style environment for every single qubit, we can automate the procedure, lessening the number of estimations and qubits required, which accelerates the entire procedure.”

Efforts have been made as of late to illuminate old-style automated tasks utilizing AI as another option. In the quantum domain, quantum neural networks could help take care of issues related with kinematics, or the mechanical movement of robots.

There are reports that state how the two degrees of control in robotics, abstract task-planning, and specific movement-planning which are presently illuminated independently, can be explained in an increasingly integrative way with quantum computing.

Quantum computing could play an important job in enhancing the development of machines, including identifying moments of inertia and joint friction. Such difficulties could be addressed with quantum reinforcement learning, with models that can develop themselves, and with “hybrid quantum-classical algorithms.”

The post THE INTER-DEPENDENCE OF QUANTUM COMPUTING AND ROBOTICS appeared first on Artificial Intelligence.

To effectively interact with their surrounding environment, robots should be able to identify characteristics of different objects just by touching them, like humans do. This would allow them to get hold of and manage objects more efficiently, using feedback gathered by sensors to adjust their grasp and manipulation strategies.

With this in mind, research groups worldwide have been trying to develop techniques that could give robots a sense of touch by analyzing data collected by sensors, many of which are based on the use of deep learning architectures. While some of these methods are promising, they typically require vast amounts of training data and do not always generalize well across previously unseen objects.

Researchers at ETH Zurich have recently introduced a new deep learning-based strategy that could enable tactile sensing in robots without requiring large amounts of real-world data. Their approach, outlined in a paper pre-published on arXiv, entails training deep neural networks entirely on simulation data.

“Our technique learns from data how to predict the distribution of the forces exerted by an object in contact with the sensing surface,” Carlo Sferrazza, one of the researchers who carried out the study, told TechXplore. “So far, this data (in the order of tens of thousands of data points) needed to be collected in an experimental setup over several hours, which was expensive in terms of time and equipment. In this work, we generated our data entirely in simulation, retaining high sensing accuracy when deploying our technique in the real world.”

In their experiments, Sferrazza and his colleagues used a sensor they built with simple and low-cost components. This sensor is comprised of a standard camera placed below a soft material, which contains a random spread of tiny plastic particles.

When a force is applied to its surface, the soft material deforms and causes the plastic particles to move. This motion is then captured by the sensor’s camera and recorded.

“We exploit the image patterns created by the moving particles to extract information about the forces causing the material deformation,” Sferrazza explained. “By densely embedding the particles into the material we can obtain an extremely high resolution. Since we take a data-driven approach to solve this task, we can overcome the complexity of modeling contact with soft materials and estimate the distribution of these forces with high accuracy.”

Essentially, the researchers created models of the sensor’s soft material and camera projection using state-of-the-art computational methods. They then used these models in simulations, to create a dataset of 13,448 synthetic images that is ideal for training tactile sensing algorithms. The fact that they were able to generate training data for their tactile sensing model in simulations is highly advantageous, as it prevented them from having to collect and annotate data in the real world.

“We also developed a transfer learning technique that allows us to use the same model on multiple instances of the tactile sensors we produce in the real-world, without the need for additional data,” Sferrazza said. “This means that each sensor becomes cheaper to produce, as they don’t require additional calibration efforts.”

The researchers used the synthetic dataset they created to train a neural network architecture for vision-based tactile sensing applications and then evaluated its performance in a series of tests. The neural network achieved remarkable results, making accurate sensing predictions on real data, even if it was trained on simulations.

“The tailored neural network architecture that we trained also shows very promising generalization possibilities for use in other situations, when applied to data that is quite different from that used in our simulations, e.g., for the estimation of contact with single or multiple objects of arbitrary shapes,” Sferrazza said.

In the future, the deep learning architecture developed by Sferrazza and his colleagues could provide robots with an artificial sense of touch, potentially enhancing their grasping and manipulation skills. In addition, the synthetic dataset they compiled could be used to train other models for tactile sensing or may inspire the creation of new simulation-based datasets.

“We now want to evaluate our algorithms in tasks that involve very general interactions with complex objects, and we are also working on improving their accuracy,” Sferrazza said. “We think that this technique will show its advantages when applied to real-world robotic tasks, such as applications that involve the fine manipulation of fragile objects—such as a glass or an egg.”

The post A deep learning-based method for vision-based tactile sensing appeared first on Artificial Intelligence.

Human intelligence is the quality of brain that learns, extracts knowledge, acquires abstract concepts from its surrounding, whereas artificial intelligence is the ability of a machine to mimic the same tasks learning from data it receives. Intelligence is a quality that belongs to humans and if machines could play the game right, our lives would become much easier.

Timo Elliott, Innovation Evangelist, SAP said, “The rise of artificial intelligence is raising the premium on tasks that only humans can do: it is freeing workers from drudgery and allowing them to spend time on more strategic and valuable business activities. Instead of forcing people to spend time and effort on tasks that we find hard but computers find easy, we will be rewarded for doing what humans do best — and artificial intelligence will help make us all more human.”

However, despite significant advancements, AI still could not match up to human intelligence in most aspects. In the growing debate about AI vs. human intelligence, the given wisdom has been that artificial intelligence will augment human tasks, but not replace them, anytime soon. Andrew McAfee, a professor at Massachusetts Institute of Technology, noted that 20 years have passed since a computer beat world chess champion Garry Kasparov yet the gap between computer ability and human ability has only gotten more significant. He said, “We still underestimate how big, how fast, technological progress is. I still keep getting it wrong.”

In just the past two years, McAfee said, AI has defied expectations.

“Certainly AI is proving to be an invaluable tool, and intelligent workflow is going to be the labor-saving norm within just a few years,” said Scott Robinson, a SharePoint and business intelligence expert based in Louisville, Ky. “But business processes involve intelligent thought and intelligent behavior. AI is great at replicating intelligent behavior, but intelligent thought is another matter. We don’t fully understand how intelligent human thoughts develop, so we’re not going to build machines that can have them anytime soon.”

“[McAfee’s] discussion misses the fact that human workers bring deep knowledge to business processes that AI can’t capture,” Robinson continued. “An office worker knows how other human beings think and behave, so she can anticipate delays or opportunities. There are implicit tasks in all areas of business that are undocumented but natural and deeply ingrained. AI can’t get anywhere near those implicit tasks and passive knowledge.”

Moreover, a book ‘The Globotics Upheaval’ by Richard Baldwin suggests that AI will disrupt lives more than globalization, industrialization, and automation did. While he believes that the changes are inevitable, there are adaptive strategies that can be used, employing the skills that no machine can copy; creativity and independent thought.

Analyzing the unbridgeable gap between human intelligence and artificial intelligence in the near future, the right perspective would be to the complementary attributes of AI with human intelligence. The scientific researches should be focused on developing artificial intelligence applications that could integrate with human intelligence in an effort to enhance productivity within the broad restrictions of privacy and sensibilities. If we will create a collaborative world for survival for both, then it would surely open up new opportunities for many in numerous different fields.

The post WILL ARTIFICIAL INTELLIGENCE GROW BEYOND HUMAN INTELLIGENCE? appeared first on Artificial Intelligence.

Artificial Intelligence will have a great impact on many aspects of our life, including IP protection. Investments on AI development are increasing and strong competition amongst various countries already started, with USA and China leading the path. EU is also planning investments and study researches, including a call for tender relating to “Trends and Developments in Artificial Intelligence – Challenges to the Intellectual Property Rights Framework” launched on March 2019. As we will see, current copyright laws in EU are probably not covering AI results. The question is whether protection is needed; and in this case what kind of protection would be more advisable, taking into consideration the effects – and the possible counter-effects – on human authors’ protection and the economy.

     1. Results produced by AI

Examples of works of art created by AI are now numerous. One case concerns the Collective “Obvious”, based in Paris, that used the IT system GAN (Generative Adversarial Network) to produce a portrait in a pictorial style between the 14th and 20th centuries. The goal was to produce a portrait that could not be distinguished by a work created by a human author, with reference to style and subject, and at the same time “original” (in the sense that it was not identical to prior paintings). The result was a work of art named “Compte de Bellamy”, that went to auction at Christies in October 2018. Similar examples relate to the songs realized by the “Creativity Machine” of Dr. Thaler, that in one single weekend produced 11,000 new music tracks. In these cases, the natural person/s collect the data input and expose the IT systems to such data, choosing the goal of the activity, however without making any decision as to the actual expressive form of the final work of art. It is the IT system that through a generator elaborates various results; such results are then selected by a second device (the discriminator) based on their indistinguishability from a work of human creation and on the basis of their “originality”.

     2. Creation in Copyright

Works of art are protected since they are the author’s own intellectual creation, i.e. the result of the choice, sequence and combination realized by the author (see ECJ 16th July 2009, C-5/08 – Infopaq). As indicated above, AI works cannot be distinguished by human creations, is so far as they look like them. Therefore, in judging from the features of the work, there should be no difference between AI and human creations, also considering that the threshold for the protection of the latter is generally considered to be rather low. However, there is a key question here to be answered, i.e. whether creation as such implies conscience and will. If this is the case, AI works could not qualify for the protection, as the IT systems could hardly be said to exercise conscience and will for the creation.

     3. Author as a natural person

Another issue concerns the possibility for a non-human entity to acquire the copyright on a work of art. According to the vast majority of copyright laws in the world, only natural persons can be authors, and acquire both economic as well as moral rights. The latter can hardly be conferred to a device. In this respect one could make reference to the debate originated in the USA in relation to the famous “Naruto” portrait, that was a photograph taken by a macaque, using the camera let on purpose unattended by the owner, the photographer David Slater, to allow spontaneous activities by the group of monkeys that he was observing in 2011. The photographer claimed to be the owner of the photograph, however the United States Court of Appeals of the Ninth Circuit found differently with its decision of April 23, 2018. In 2014 the US Copyright Office expressly established that copyright can only protect original works of authorship, i.e. created by a human being, excluding works created by “nature, animals, or plants”. Therefore, it would seem that AI works autonomously produced by the IT system could not qualify for copyright protection. In other words, protection could be granted only when there is human intervention in the process carried out by the AI. After all, one has to bear in mind that the approach toward AI should be based on the key principle according to which the center and focus of the protection is and remain the human being. Therefore, AI is a tool and should not be the goal or the focus of the protection.

     4. If AI cannot be protected by copyright, is there a need for another form of protection?

If AI cannot be protected by copyright, as indicated above, the fear exists that this will result in decrease of investments in AI, thus limiting the possibility of quickly reaching all the positive effects that AI seems to promise. For this reason, some countries are going far and debating the possibility and opportunity to even granting legal subjectivity to AI (for instance, Saudi Arabia would have granted citizenship to a robot named Sophie, while Tokyo would have granted residence to another robot called Shibuya Mirai). Setting aside these extreme positions, that would require careful examination of many issues (first of all, ethical and philosophic issues), it would seem that there are a number of options available to try and protect AI’s results. For instance, one could reconsider the notion of creativity, and extend it to comprehend works created by AI, enhancing human intervention consisting of selecting the data collected and entered into the machine and choosing the parameters that define the objective of the machine’s activity. It is also possible to think of attributing a kind of sui generis right to AI (rather than copyright), also given the circumstance that the value behind AI seems to concern more the investments than the creativity.

     5. “Killing” human authors?

Finally, it has to be considered that AI produces (and will increasingly produce) high amounts of works that will not be distinguishable from works created by human beings, as indicated above. One certainly shares the view that AI should be used to improve wellbeing for human beings, rather than the other way around. Therefore, it would seem rather reasonable to try and maintain high protection for human beings when they create works of art, and exclude any protection, or grant minor protection, to the results of AI activities. However, it is necessary to also think to possible countereffects of this approach. Works created by human beings, albeit substantially similar to those created by AI, will present higher barriers to exploitation than the latter, that would be in public domain, or subject to a shorter deadline for the protection, or in any event probably cheaper. Would this benefit and promote human creativity, or will there be an opposite effect? In this scenario, a possibility is that the notion of creation (and the relating copyright protection) is limited to those works of art realized by a human being, which are the result of an intuitive creative act, not reproducible by a machine. However, since creativity and its nature has been quite a complex issue to define (and still is), one wonders whether it is actually possible to draw the line above indicated. Also, it should be carefully consider how this conclusion will impact to the job market in the creative industries (setting aside the circumstance that AI will probably dramatically change the way through which such industries will operate).

The post Artificial Intelligence and Copyright Protection appeared first on Artificial Intelligence.