What are neural-symbolic AI methods and why will they dominate 2020? – The Next Web
Posted: January 18, 2020 at 4:42 pm
The recent commercial AI revolution has been largely driven by deep neural networks. First invented in the 1960s, deep NNs came into their own once fueled by the combination of internet-scale datasets and distributed GPU farms.
But the field of AI is much richer than just this one type of algorithm. Symbolic reasoning algorithms such as artificial logic systems, also pioneered in the 60s, may be poised to emerge into the spotlight to some extent perhaps on their own, but also hybridized with neural networks in the form of so-called neural-symbolic systems.
Deep neural nets have done amazing things for certain tasks, such as image recognition and machine translation. However, for many more complex applications, traditional deep learning approaches cannot match the ability of hybrid architecture systems that additionally leverage other AI techniques such as probabilistic reasoning, seed ontologies, and self-reprogramming ability.
Deep neural networks, by themselves, lack strong generalization, i.e. discovering new regularities and extrapolating beyond training sets. Deep neural networks interpolate and approximate on what is already known, which is why they cannot truly be creative in the sense that humans can, though they can produce creative-looking works that vary on the data they have ingested.
This is why large training sets are required to teach deep neural networks and also why data augmentation is such an important technique for deep learning, which needs humans to specify known data transformations. Even interpolation cannot be done perfectly without learning underlying regularities, which is vividly demonstrated by well-known adversarial attacks on deep neural networks.
The slavish adherence of deep neural nets to the particulars of their training data also makes them poorly interpretable. Humans cannot completely rely or interpret their results, especially in novel situations.
What is interesting is that, for the most part, the disadvantages of deep neural nets are strengths of symbolic systems (and vice versa), which inherently possess compositionality, interpretability, and can exhibit true generalization. Prior knowledge can also be easily incorporated into symbolic systems in contrast to neural nets.
Neural net architectures are very powerful at certain types of learning, modeling, and action but have limited capability for abstraction. That is why they are compared with the Ptolemaic epicycle model of our solar system they can become more and more precise, but they need more and more parameters and data for this, and they, by themselves, cannot discover Keplers laws and incorporate them into the knowledge base, and further infer Newtons laws from them.
Symbolic AI is powerful at manipulating and modeling abstractions, but deals poorly with massive empirical data streams.
This is why we believe that deep integration of neural and symbolic AI systems is the most viable path to human-level AGI on modern computer hardware.
Its worth noting in this light that many recent deep neural net successes are actually hybrid architectures, e.g. the AlphaGo architecture from Google DeepMind integrates two neural nets with one game tree. Their recent MuZero architecture, which can master both board and Atari games, goes further along this path using deep neural nets together with planning with a learned model.
The highly successful ERNIE architecture for Natural Language Processing question-answering from Tsinghua University integrates knowledge graphs into neural networks. The symbolic sides of these particular architectures are relatively simplistic, but they can be seen as pointing in the direction of more sophisticated neural-symbolic hybrid systems.
The integration of neural and symbolic methods relies heavily on what has been the most profound revolution in AI in the last 20 years the rise of probabilistic methods: e.g. neural generative models, Bayesian inference techniques, estimation of distribution algorithms, probabilistic programming.
As an example of the emerging practical applications of probabilistic neural-symbolic methods, at the Artificial General Intelligence (AGI) 2019 conference in Shenzhen last August, Hugo Latapie from Cisco Systems described work his team has done in collaboration with our AI team at SingularityNET Foundation, using the OpenCog AGI engine together with deep neural networks to analyze street scenes.
The OpenCog framework provides a neural-symbolic framework that is especially rich on the symbolic side, and interoperates with popular deep neural net frameworks. It features a combination of probabilistic logic networks (PLNs), probabilistic evolutionary program learning (MOSES), and probabilistic generative neural networks.
The traffic analytics system demonstrated by Latapie deploys OpenCog-based symbolic reasoning on top of deep neural models for street scene cameras, enabling feats such as semantic anomaly detection (flagging collisions, jaywalking, and other deviations from expectation), unsupervised scene labeling for new cameras, and single-shot transfer learning (e.g. learning about new signals for bus stops with a single example).
The difference between a pure deep neural net approach and a neural-symbolic approach in this case is stark. With deep neural nets deployed in a straightforward way, each neural network models what is seen by a single camera. Forming a holistic view of whats happening at a given intersection, let alone across a whole city, is much more of a challenge.
In the neural-symbolic architecture, the symbolic layer provides a shared ontology, so all cameras can be connected for to an integrated traffic management system. If an ambulance needs to be routed in a way that will neither encounter nor cause significant traffic, this sort of whole-scenario symbolic understanding is exactly what one needs.
The same architecture can be applied to many other related use cases where one can use neural-symbolic AI to both enrich local intelligence and connect multiple sources/locations into a holistic view for reasoning and action.
It may not be impossible to crack this particular problem using a more complex deep neural net architecture, with multiple neural nets working together in subtle ways. However, this is an example of something that is easier and more straightforward to address using a neural-symbolic approach. And it is quite close to machine vision, one of deep neural nets great strengths.
In other, more abstract application domains such as mathematical theorem-proving or biomedical discovery the critical value of the symbolic side of the neural-symbolic hybrid is even more dramatic.
Deep neural nets have done amazing things over the last few years, bringing applied AI to a whole new level. Were betting that the next phase of incredible AI achievements are going to be delivered via hybrid AI architectures such as neural-symbolic systems. This trend has already started in 2019 in a relatively quiet way and in 2020 we expect it will pick up speed dramatically.
Published January 15, 2020 09:00 UTC
The rest is here:
What are neural-symbolic AI methods and why will they dominate 2020? - The Next Web
- Facebooks Hanabi-playing AI achieves state-of-the-art results - VentureBeat [Last Updated On: December 11th, 2019] [Originally Added On: December 11th, 2019]
- Biggest scientific discoveries of the 2010s decade: photos - Business Insider [Last Updated On: December 11th, 2019] [Originally Added On: December 11th, 2019]
- DeepMind co-founder moves to Google as the AI lab positions itself for the future - The Verge [Last Updated On: December 11th, 2019] [Originally Added On: December 11th, 2019]
- AlphaGo - Wikipedia [Last Updated On: December 11th, 2019] [Originally Added On: December 11th, 2019]
- DeepMind Vs Google: The Inner Feud Between Two Tech Behemoths - Analytics India Magazine [Last Updated On: December 18th, 2019] [Originally Added On: December 18th, 2019]
- AI is dangerous, but not for the reasons you think. - OUPblog [Last Updated On: December 18th, 2019] [Originally Added On: December 18th, 2019]
- The Perils and Promise of Artificial Conscientiousness - WIRED [Last Updated On: December 18th, 2019] [Originally Added On: December 18th, 2019]
- AI has bested chess and Go, but it struggles to find a diamond in Minecraft - The Verge [Last Updated On: December 18th, 2019] [Originally Added On: December 18th, 2019]
- What is AlphaGo? - Definition from WhatIs.com [Last Updated On: December 22nd, 2019] [Originally Added On: December 22nd, 2019]
- AlphaZero beat humans at Chess and StarCraft, now it's working with quantum computers - The Next Web [Last Updated On: January 18th, 2020] [Originally Added On: January 18th, 2020]
- Why asking an AI to explain itself can make things worse - MIT Technology Review [Last Updated On: January 29th, 2020] [Originally Added On: January 29th, 2020]
- Why The Race For AI Dominance Is More Global Than You Think - Forbes [Last Updated On: February 10th, 2020] [Originally Added On: February 10th, 2020]
- AI on steroids: Much bigger neural nets to come with new hardware, say Bengio, Hinton, and LeCun - ZDNet [Last Updated On: February 10th, 2020] [Originally Added On: February 10th, 2020]
- I think, therefore I am said the machine to the stunned humans - Innovation Excellence [Last Updated On: February 10th, 2020] [Originally Added On: February 10th, 2020]
- From Deception to Attrition: AI and the Changing Face of Warfare - War on the Rocks [Last Updated On: February 20th, 2020] [Originally Added On: February 20th, 2020]
- Levels And Limits Of AI - Forbes [Last Updated On: February 20th, 2020] [Originally Added On: February 20th, 2020]
- How to overcome the limitations of AI - TechTarget [Last Updated On: February 20th, 2020] [Originally Added On: February 20th, 2020]
- The top 5 technologies that will change health care over the next decade - MarketWatch [Last Updated On: February 25th, 2020] [Originally Added On: February 25th, 2020]
- Chess grandmaster Gary Kasparov predicts AI will disrupt 96 percent of all jobs - The Next Web [Last Updated On: February 25th, 2020] [Originally Added On: February 25th, 2020]
- Enterprise AI Books to Read This Spring - DevOps.com [Last Updated On: March 14th, 2020] [Originally Added On: March 14th, 2020]
- The New ABCs: Artificial Intelligence, Blockchain And How Each Complements The Other - JD Supra [Last Updated On: March 14th, 2020] [Originally Added On: March 14th, 2020]
- The Turing Test is Dead. Long Live The Lovelace Test - Walter Bradley Center for Natural and Artificial Intelligence [Last Updated On: April 8th, 2020] [Originally Added On: April 8th, 2020]
- QuickBooks is still the gold standard for small business accounting. Learn how it's done now. - The Next Web [Last Updated On: April 19th, 2020] [Originally Added On: April 19th, 2020]
- This A.I. makes up gibberish words and definitions that sound astonishingly real - Digital Trends [Last Updated On: May 17th, 2020] [Originally Added On: May 17th, 2020]
- The Hardware in Microsofts OpenAI Supercomputer Is Insane - ENGINEERING.com [Last Updated On: June 5th, 2020] [Originally Added On: June 5th, 2020]
- Why the buzz around DeepMind is dissipating as it transitions from games to science - CNBC [Last Updated On: June 5th, 2020] [Originally Added On: June 5th, 2020]
- AlphaGo (2017) - Rotten Tomatoes [Last Updated On: June 5th, 2020] [Originally Added On: June 5th, 2020]
- AlphaGo - Top Documentary Films [Last Updated On: June 5th, 2020] [Originally Added On: June 5th, 2020]
- Enterprise hits and misses - contactless payments on the rise, equality on the corporate agenda, and Zoom and Slack in review - Diginomica [Last Updated On: June 8th, 2020] [Originally Added On: June 8th, 2020]
- Is Dystopian Future Inevitable with Unprecedented Advancements in AI? - Analytics Insight [Last Updated On: June 26th, 2020] [Originally Added On: June 26th, 2020]
- Test your Python skills with these 10 projects - Best gaming pro [Last Updated On: October 3rd, 2020] [Originally Added On: October 3rd, 2020]
- In the Know - UCI News [Last Updated On: October 3rd, 2020] [Originally Added On: October 3rd, 2020]
- How to Understand if AI is Swapping Civilization - Analytics Insight [Last Updated On: October 3rd, 2020] [Originally Added On: October 3rd, 2020]
- Investing in Artificial Intelligence (AI) - Everything You Need to Know - Securities.io [Last Updated On: November 2nd, 2020] [Originally Added On: November 2nd, 2020]
- What the hell is reinforcement learning and how does it work? - The Next Web [Last Updated On: November 2nd, 2020] [Originally Added On: November 2nd, 2020]
- An AI winter may be inevitable. What we should fear more: an AI ice age - ITProPortal [Last Updated On: December 4th, 2020] [Originally Added On: December 4th, 2020]
- Are Computers That Win at Chess Smarter Than Geniuses? - Walter Bradley Center for Natural and Artificial Intelligence [Last Updated On: December 4th, 2020] [Originally Added On: December 4th, 2020]
- What are proteins and why do they fold? - DW (English) [Last Updated On: December 12th, 2020] [Originally Added On: December 12th, 2020]
- Are we ready for bots with feelings? Life Hacks by Charles Assisi - Hindustan Times [Last Updated On: December 12th, 2020] [Originally Added On: December 12th, 2020]
- Examining the world through signals and systems - MIT News [Last Updated On: February 10th, 2021] [Originally Added On: February 10th, 2021]
- How AI is being used for COVID-19 vaccine creation and distribution - TechRepublic [Last Updated On: April 24th, 2021] [Originally Added On: April 24th, 2021]
- The 13 Best Deep Learning Courses and Online Training for 2021 - Solutions Review [Last Updated On: April 24th, 2021] [Originally Added On: April 24th, 2021]
- Why AI That Teaches Itself to Achieve a Goal Is the Next Big Thing - Harvard Business Review [Last Updated On: April 24th, 2021] [Originally Added On: April 24th, 2021]
- The Alpha of 'Go'. What is AlphaGo? | by Christopher Golizio | Apr, 2021 | Medium - Medium [Last Updated On: April 24th, 2021] [Originally Added On: April 24th, 2021]
- How will Edge Artificial Intelligence (AI) Chips Take IoT Devices to the Next Level - Enterprise Apps Today [Last Updated On: July 6th, 2022] [Originally Added On: July 6th, 2022]
- Machines with Minds? The Lovelace Test vs. the Turing Test - Walter Bradley Center for Natural and Artificial Intelligence [Last Updated On: July 6th, 2022] [Originally Added On: July 6th, 2022]
- For AI to Be Creative, Here's What It Would Take - Discovery Institute [Last Updated On: July 6th, 2022] [Originally Added On: July 6th, 2022]
- What is my chatbot thinking? Nothing. Here's why the Google sentient bot debate is flawed - Diginomica [Last Updated On: August 7th, 2022] [Originally Added On: August 7th, 2022]
- Incoherent, creepy and gorgeous: we asked six leading artists to make work using AI and here are the results - The Guardian [Last Updated On: December 4th, 2022] [Originally Added On: December 4th, 2022]
- Top 5 Applications of Reinforcement Learning in Real-Life - Analytics Insight [Last Updated On: December 4th, 2022] [Originally Added On: December 4th, 2022]
- OpenAI tweaks ChatGPT to avoid dangerous AI information - The Register [Last Updated On: December 4th, 2022] [Originally Added On: December 4th, 2022]
- Go champion who faced off against Google's AlphaGo says the rise of AI strips the games of artistry - DIGITIMES [Last Updated On: April 4th, 2024] [Originally Added On: April 4th, 2024]