Seeflection is bringing primary AI research together with Quantum Computing across several critical areas. These include mitigating quantum noise impact on calculations through use of adaptive AI input and Deep Learning stages, in addition to researching the use of parallel quantum results for areas including optimal hyper-parameter selection.
We are applying decades of primary and applied research and development by some of our key team members to interdisciplinary quantum and classical challenges. We are also leveraging internal and external resources to develop new optimization and problem-solving approaches.
In areas related to applying Artificial Intelligence to Quantum Computing challenges, we are researching Deep Learning-enabled information recovery and reconstruction architectures to effectively reach below the Shannon’s Limit noise floor.
We are researching the application Quantum Computing’s potential algorithmic advantages to Artificial Intelligence challenges, and are exploring methods for selection of optimal initial hyper-parameters and adaptive parameter states. A related area of research includes creating quantum-classical bias and weight calculation methods as alternatives to back and forward propagation methods.
Hyper-parameters – Identifying optimal initial set, and active adaptation
Applied Deep Learning – leveraging AI to mitigate quantum noise
Quantum Optimization – Identification of optimal global and local minima/maxima starting points
Convolution and Deconvolution – Quantum computing enabled methods in Deep Learning models
Quantum Computing Challenges – Apply AI to address Quantum Computing challenges
Artificial Intelligence – Use Quantum Computing to address AI limitations
Quantum-AI – Create qualitatively new and better combined solutions
Quantum Computing AI Lab
AI is uniquely able to address a huge number of challenges, such as:
- Enabling doctors to diagnose conditions more efficiently and accurately than humanly possible.
- Quantum computer research is already focusing on solving problems like:
- A wide range of scientific, transportation, workflow and other types of optimization problems.
There are even more areas that are not yet addressable by either AI or Quantum computing. Part of this is due to the fact that AI is only now exploding in a practical, commercial sense, since its widespread usability just a few years ago.
Similarly, commercially accessible Quantum Computers have only just become available in the past few years as well. It has demonstrated unique advantages already in optimization solutions, which interestingly is what most Artificial Intelligence applications are at their core.
The combination of AI and Quantum Computing is a highly complementary one, where the result becomes substantially more than the sum of its parts.
Optimization Problem & Solutions
The evolutionary timeline of Quantum Computing (QC), Artificial Intelligence (AI) and Classical Computing (CC), appears almost as if it’s based on some external plan. Seeflection’s research is positioned to help leverage these trends, while working to optimize the benefits to society and commerce.
Many of the current designs of Quantum Computers, such as those from D-Wave, are optimal for solving optimization problem; which interestingly, are what many of the current types of Artificial Intelligence learners and system-level solutions are based around solving.
While AI is rapidly accelerating towards general purpose, or “Artificial General Intelligence” or AGI over the next several, Quantum Computer researchers are working on general purpose Quantum Computing, which is projected to be developed over about the same time-frame of the next several years.
During the same projected period, Moore’s Law that state that the price/performance of classical computers doubles every 18 months is projected to finally
At the same time that classical computing’s growth is leveling-off, Quantum Computing and AI are about to take over as the dominant digital solution paradigm. Although traditional computers have a critical role in the ongoing digital landscape, Concurrently, the growth rate of Moore’s law, as well as the types of problems that are possible to solve, are about to undergo a radical shift.
Primary examples of this shift include:
A growth rate that appears will likely accelerate from doubling every 18 months, to a projected 700-fold or grater increase every 12 months.
Recursively self-improving AI with effectively unlimited growth rate.
In addition to working on projects to leverage those trends, Seeflection is leveraging emergent capabilities, which already seem to be a dominant part of the combined evolution of AI and the new digital landscape. Seeflection has begun researching ways to leverage emergent behaviors, such as autonomous language and problem-solving capabilities resulting from:
Nonlinearities, disjunctions and/or bias being introduced to convolved and de-convolved areas. This is analogous to neurology, where emergent cognitive abilities result from the introduction of changes to white matter interconnectivity or the bias between areas, such as from changes to physical brain structures.
A threshold amount of new contiguous information added or taken away, while there is a new connectivity bias introduced affecting the method for converging on a new classification, regression solution, or pooling or propagation activities.
New Convergent Capabilities – leveraging the combined capabilities of AI, Quantum Computing and big data
New Emergent Capabilities – resulting from autonomous adjustment and adaptation
Growth related needs – from accelerated introductions of new and disruptive capabilities
Medical Research – Protein folding quantum-classical solutions
Transportation/Transport – Challenges and solutions affecting society
Health – New ways of addressing issues and delivering care
In the past 5 years, AI has come further in terms of practical applicability than in the prior 50 years since its inception. Thanks to dramatic increases in computer hardware available to AI, coupled with some amazing human research, AI is already delivering on its promise to offer dramatically better ways of solving problems and providing solutions.
However, as amazing as the promise of Artificial intelligence is; when combined with Quantum Computing through applied human insight and ingenuity, it promises to be nothing short of miraculous. We are among those who seek to inform, educate and enable that revolution.
The Future – Quantum-AI
AI is transforming Big data from big headaches to huge benefits– especially in the area of Deep Learning Neural Networks. The availability of Big Data is enabling AI to become self-learning; replacing tedious and expensive manual expert system related knowledge capture, restructuring, and ingestion into a manually coded rule-based architecture. This has come with huge benefirs, including:
- The time and cost of creating new Learners has be reduced by orders of magnitude
- The accuracy, efficiency and performance of AI on a wide range of specific application has gone from proof of concept to super-human virtually overnight.
- The types of problems that can be solved, and the inferences obtained have exploded
The Internet has formed the backbone for access and delivery to much of this. However, while huge amounts of data have become readily available, new opportunities have emerged in many new areas, including:
- Intelligent data and media search, ingestion, aggregation, management, production and post-production.
- Automated news and information research and validation testing
- Fundamental research, proofs, corollaries and
- Autonomous inductive knowledge acquisition and validation
See AI Reflected in your Future
We are a multi-disciplinary research center committed to solving real world problems using leading-edge technologies.