Thinkers360
Interested in getting your own thought leader profile? Get Started Today.

Scot Forshaw

Research Director at Toridion Project

Manchester, United Kingdom

I am interested in the fundamental study of information, turbulence, scale and transformability, particularly related to how quantum algorithms such as quantum annealing can be harnessed to generate intelligent neural networks for universal deep learning systems. My research crosses into neurological sciences and consciousness studies, particularly the study and development of macromolecular simulations to emulate brain like memory storage and retrieval in finite simulated quantum systems.

Scot ForshawPoints
Academic35
Author8
Influencer46
Speaker0
Entrepreneur0
Total89

Points based upon Thinkers360 patent-pending algorithm.

Thought Leader Profile

Portfolio Mix

Company Information

Company Type: Company
Business Unit: AI and Quantum Neural Networks
Theatre: Europe, ASEAN

Areas of Expertise

Agile
AI 32.38
Big Data 30.85
Cloud
Cryptocurrency 33.85
Cybersecurity
Design Thinking
Emerging Technology
ERP
Govtech
Healthtech
IoT
Predictive Analytics 33.71
Quantum Computing 100
Climate Change 30.58

Industry Experience

Engineering & Construction
Federal & Public Sector
Financial Services & Banking
High Tech & Electronics
Industrial Machinery & Components
Retail
Telecommunications

Please signin or signup to view publication section.

Publications

5 Article/Blogs
The Artificial Intelligence Race won't be won by GPU's and TFLOPS
Linkedin
August 19, 2020
As a recent MIT/IBM joint paper offers the strongest evidence yet that AI and Deep Learning are facing a performance crisis[1], I have to look critically at the big names still busy pushing away at GPU and logic based incremental improvements. The maths is in... it isn't going to work! Simply stuffing billions more transistors in a die and burning more energy won't reach the terminal velocity required to accelerate Deep Learning beyond a few more iterations of performance increase.

See publication

Tags: AI, Predictive Analytics, Quantum Computing

QUAKESCANNER MISSION I
Visicom Scientific
December 31, 2016
All the latest data, forecasts and mission results from the Quakescanner Earthquake I Mission! Earthquakes are one of the most powerful and destructive natural forces on earth. Every year they claim 100's if not 1000's of lives globally. Forecasting earthquakes is notoriously difficult and despite millions of dollars of research annually, the ability to forecast them is still far from a reality.

See publication

Tags: Climate Change, Predictive Analytics

An Artificial Life Developers Response to Deepak Chopra's Article - “Artificial Intelligence Will Never Rival the Deep Complexity of the Human Mind”
Linkedin
July 05, 2016
“AI isn’t based on the truth.”

The article opens with the above assertion ...

As an AL developer and researcher I have no objection to the foregoing statement. I agree with Deepak Chopra, and would say “reality is quite obviously based upon uncertainty and principles that cannot be readily and recursively expressed or fully calculated – therefore very little is 'TRUE' in the scientific or philosophical sense of the word”. I cover this subject in much more detail in my paper “The Third State: Toward a Quantum Information Theory of Consciousness” (Forshaw, 2015).

See publication

Tags: AI

Question Generation Technology (QGT) in Deep Learning - Why Defining New Questions Is More Important Than Finding Answers in Data Analytics.
Linkedin
November 04, 2015
For anyone heavily involved in the development of Deep Learning technology, by far one of the most difficult things we face is actually putting what is in our heads into meaningful words. By this I mean answering the many questions that the inquisitive audience ask of us. For example, I may be asked “what is this hidden region then?”.

See publication

Tags: AI, Predictive Analytics, Quantum Computing

Neuroplasticity demonstrated in a Zero Logic Quantum Neural Network
Linkedin
August 02, 2015
A recent development at Visicom Scientific is the creation of a primitive but highly capable Zero Logic Quantum Probabilistic Neural Network. Or ZLQNN for short, and in recent experiments (highlighted below), it was shown how information stored in a probabilistic form of layered frequencies could be shown to not only store information with 100% accuracy but also self organise, self repair and self create useful information without the use of logic or dedicated software to direct it.

See publication

Tags: AI, Predictive Analytics, Quantum Computing

1 Conference Publication
Re-constructing memory using quantized electronic music and a ‘Toridion byte’ quantum algorithm: Creating images using zero logic quantum probabilistic neural networks (ZLQNN)
Intellect
December 20, 2016
Quantum theory applied to data analytics using a quantum computer has become the leading research endeavor to find a way to store and retrieve data using the nano-sized world of molecular structures. The ability to manipulate and change the positioning of magnetic fields in the sub atomic realms through resonant frequencies has provided new insights into how a probabilistic and indeterminate field of energy (quantum) can be utilized for storing and retrieving data. Much of the theorization that is applied to quantum computer development relies on a conceptual framework largely based on metaphors to understand the behavior of sub-atomic elements within a quantum field (Brookes, 2014). One aspect of the quantum field is the entanglement of elements whereby behaviors of two distinct elements respond to change independent of their location. The Toridion quantum algorithm was used to scatter pre-recorded sound into frequency amplitudes within a simulated quantum computer environment. The sounds were composed by using quantum cognitive meta models for the creation of electronic music compositions. The Toridion Encoder creates highly compressed ‘glyphs' of the sounds whilst simultaneously creating a probabilistic quantum neural network within the cyclic mental workspace of the computer. By using sensory experiences as non deterministic search functions, it is explored how a quantum machine learning algorithm is able to unlock images, sounds, text or other media that is compressed into these small data packets called a “Toridion byte”. Because of the super positioning effect of a quantized state, in which a state can be in several locations at once, a reduction in size of the stored data is possible by using this compression technique. The compression of the data is further reduced until all that is left is a single packet of data within the quantum neural network. Images of prior events that are compressed and stored in a simulated quantum computer environment are then retrieved using encoded elements from the pre-recorded sound frequencies. This paper will explain how using a quantum compositional framework in composing electronic music orchestrations can aid in retrieving

(PDF) Re-constructing memory using quantized electronic music and a “Toridion byte” quantum algorithm: Creating images using zero logic quantum probabilistic neural networks (ZLQNN). Available from: https://www.researchgate.net/publication/286452191_Re-constructing_memory_using_quantized_electronic_music_and_a_Toridion_byte_quantum_algorithm_Creating_images_using_zero_logic_quantum_probabilistic_neural_networks_ZLQNN [accessed Feb 20 2020].

See publication

Tags: AI, Predictive Analytics, Quantum Computing

6 Journal Publications
Toridion QML API BTC (Bitcoin) Cryptocurrency Prediction Performance 37 Hr Test October 1st - 2nd 2020 Results with Data
Toridion
October 03, 2020
We review the performance of the live predictions made by our API for the period 10/01/2020 9am UTC to 10/02/2020 9pm inclusive. The overall accuracy was calculated at 56.76% with 21 / 37 correct predictions. At a requested confidence >= 50% the accuracy was 21/36 = 58.33%. At 54% requested confidence, the accuracy was 13/19 = 68.42%. These results show that the quantum inspired machine learning model was successful at predicting Bitcoin moving average direction for > 56% of all market positions tested. Furthermore, where the requested confidence of the prediction was set to 54% (considered the forex trader magic win/loss threshold[1]) accuracy was 68.42%.

See publication

Tags: Cryptocurrency, Predictive Analytics, Quantum Computing

Quantum Inspired Machine Learning for Ethereum Cryptocurrency 2020 Forward Prediction with Results
Research Gate
September 24, 2020
In this paper we present empirical results for 1 hour future prediction of the Ethereum cryptocurrency using quantum inspired machine learning [QML] for the month of January 2020. The results demonstrate accuracy that exceeds the 'magic win/loss' ratio of 54%. There is a generally held view that financial markets behave in a mostly unpredictable way[1]. Indeed observing the number of up and down events across the majority of currency pairs such as USD/GBP will reveal a largely even 50/50 distribution of events, lending weight to the random nature argument. In this study, we used a previously trained model of the Ethereum cryptocurrency[2] to forward test the models ability to confidently identify market patters with strongly probable outcomes 1 hour in the future. The model was tested for the month of January 2020 on a trained model for September 2017 to December 2019. Returning a 55-70% prediction accuracy at confidence levels >= 63%, the results demonstrate that cryptocurrency signal generation is a good candidate for quantum inspired machine learning powered predictive analytics.

See publication

Tags: Big Data, Predictive Analytics, Quantum Computing

Toridion Quantum Inspired Machine Learning for Ethereum Cryptocurrency Prediction with Results
Research Gate
September 22, 2020
There is a generally held view that financial markets behave in a mostly unpredictable way[1]. Indeed observing the number of up and down events across the majority of currency pairs such as USD/GBP will reveal a largely even 50/50 distribution of events, lending weight to the random nature argument. In this study, we applied quantum inspired machine learning to 3 years of forex and cryptocurrency market data in an attempt to build a statistically strong model for the short term forecasting of the Ethereum cryptocurrency. The machine learning model was simultaneously trained on 8 fiat currency pairs and 3 cryptocurrency pairs. Subsequent back testing of the model for various levels of confidence indicated that the model was able to predict 1 hour in the future with better than 75% accuracy when the requested confidence level was > 63%. At higher confidence levels the accuracy increased until reaching 100% accuracy at 91% confidence. The results indicate that cryptocurrency signal generation is a good candidate for quantum inspired machine learning powered predictive analytics.

See publication

Tags: Cryptocurrency, Predictive Analytics, Quantum Computing

Computational Advantages of Probabilistic Quantum Neural Networks in 'Real World' Applications
Toridion
April 20, 2017
The achievement of a useful General Purpose Quantum Neural Network [QNN] promises the delivery of an important milestone on the journey toward harnessing the effective power of quantum computing when applied to data intensive application such as big data analytics. Here, we demonstrate how Toridion Quantum Neural Networks [TQNN] can deliver significant computational performance advantages over classical computers in solving specific classes of problems with intractable “time to solution discovery” time complexities, e.g :- Toridion performs with a O(n) time complexity for a “combination of sets” matching problem such as reordering a finite 2 dimensional bitmap of n = (xy) bits with m set bits (m) and k unset bits (k) and matching this to a known pre-configuration. This performance is in stark contrast to a comparable classical memory remap and compare algorithm which devolves into near exponential time O(e) as the size and/or entropy of the bitmap set increases. Whilst we recognise that there exists classical memory sort and compare algorithms that can achieve similar results as the TQNN (in some cases), attention is drawn to the fact that the example tests are constructed to show that TQNN “learns the optimal solution” – as opposed to being “programmed” with it. It is also able to learn more than one optimal solution to more than one problem for the same dataset. This distinction is vitally important as it highlights the ability of the TQNN memory system to concurrently learn the best solution to several classes of intractable problems. This capability sets it apart as a true General Purpose Quantum Neural Network capable of enhancing the flow and processing of data at scale in real world commercial applications

See publication

Tags: AI, Predictive Analytics, Quantum Computing

The Third State: Toward a Quantum Information Theory of Consciousness
NeuroQuantology
December 15, 2015
The question of how our perceived reality is constructed and subsequently how our mind has evolved such that we are able to both perceive and subsequently alter our own causality or even our own evolution within this reality has been a long running open question. Referred to by David Chalmers[1] as “The Hard Problem”, there have been many theoretical interpretations on the nature of causal self observance – hereafter referred to as 'consciousness'. The current paper introduces the reader to the indeterminable operator - “The Third State”. The Third State is a term used to describe space itself in relation to the position of all things. As the paper shall show, The Third State is a required omnipresent and universal operator in the otherwise binary realm of data → information. The Third State augments the accepted binary operators to produce the required 'tristate' conditions that facilitates the required probabilistic nature of the conscious manifestation. Secondly the Unity Magnitude [Um] scale, which facilitates the bounding of quantum probabilistic memory in a finite model. The paper further introduces the reader to a model and experimental theory that suggests all things we perceive as physical reality can be fabricated from primitive components of data, bits – matter / antimatter / something or nothing. That facilitated by the Third State, our immediate present reality as we experience it and theretofore consciousness is a simultaneous product of the current physical configuration of the frequency stable systems of which we are comprised*, interpreted past reality, self predicted future provided by cyclic frequency stable systems and the immediate physical and sensory state including recursive imagination systems generated by the output of these perturbed frequency stable systems in the cyclic feedback process - ultimately perturbed by, but as one unified with the stochastic processes in relation to the quantum cosmological domain.

See publication

Tags: AI, Quantum Computing

Neuroplasticity demonstrated in a Zero Logic Quantum Neural Network (ZLQNN)
Toridion
September 15, 2015
The system demonstrates both simulated “neuroplasticity” and further that the fundamental act of making mistakes in self programming of a machine learning system is both a driver to finding solutions to known unknowns and developing neural pathways that can form “new opinions” that explore “unknown un

(3) (PDF) Neuroplasticity demonstrated in a Zero Logic Quantum Neural Network. Available from: https://www.researchgate.net/publication/281583936_Neuroplasticity_demonstrated_in_a_Zero_Logic_Quantum_Neural_Network [accessed Feb 20 2020].

See publication

Tags: AI, Predictive Analytics, Quantum Computing

1 Video
Inside the mind of a Quantum Artificial Intelligence
Linkedin
September 09, 2015
A short video review of our paper for the 18th DeTao Master Academy conference "Consciousness Reframed" in Shanghai - Nov 20-22 2015. A collaboration between Willard Van De Bogart of the Bangkok University Languages Institute and Scot D Forshaw a quantum algorithm and deep machine learning developer at Toridion Project.

See publication

Tags: AI, Quantum Computing

Blog

Opportunities

Contact Scot Forshaw

Media Kit

Share Profile

Contact Info

  Profile

Scot Forshaw