Artificial General Intelligence (AGI)
Artificial General Intelligence (AGI) is an evolving field aiming to build general-purpose systems with intelligence comparable to that of the human mind. What is currently labeled ‘artificial intelligence is largely narrow automated knowledge work, lacking the flexibility and adaptability seen in animal intelligence. Currently Narrow AI doesn’t have ground understanding of what it generates or what it learns, they are just statistical machines which learn from patterns in the data. The pursuit of AGI begins at a foundational level, the way humans learn, asking fundamental questions about models of cognition, knowledge representation and acquisition, making choices through reasoning, thinking and conceiving the world in adaptive and intuitive ways.
Yoshua Bengio, most known for his pioneering work in deep learning, earning him the 2018 A.M. Turing Award, “the Nobel Prize of Computing,” with Geoffrey Hinton and Yann LeCun calls current Narrow AI as system 1 deep learning, according to him ” “System 1 are the kinds of things that we do intuitively, unconsciously, that we can’t explain verbally, in the case of behavior, things that are habitual,” Bengio said. “This is what current deep learning is good at.”
Here’s how Bengio explains the difference between system 1 and system 2: Imagine driving in a familiar neighborhood. You can usually navigate the area subconsciously, using visual cues that you’ve seen hundreds of times. You don’t need to follow directions. You might even carry out a conversation with other passengers without focusing too much on your driving.
But when you move to a new area, where you don’t know the streets and the sights are new, you must focus more on the street signs, use maps and get help from other indicators to find your destination.
The latter scenario is where your system 2 cognition kicks into play. It helps humans generalize previously gained knowledge and experience to new settings. “What’s going on there is you’re generalizing in a more powerful way and you’re doing it in a conscious way that you can explain,” Bengio said at NeurIPS.
“The kinds of things we do with system 2 include programming. So we come up with algorithms, recipes, we can plan, reason, use logic,” Bengio says. “Usually, these things are very slow if you compare to what computers do for some of these problems. These are the things that we want future deep learning to do as well.”
AGI is really a subjective topic, most of the general public would have a Hollywood vision of a dystopian future. Some believe AGI needs to be self aware and conscious like humans, the reality is AGI can be a completely new intelligence which humanity is not aware of, but before we think of building conscious machines we need to solve simple problems first to fix current AI and reach strong AI/Real AI/AGI:
- Selective Focus and attention
- Shared attention, learning and teaching
- Track Prediction errors and use to update models
- Memory- working, episodic, procedural,long term
- Habituate and associate stimuli
- Replay and consolidate memories
- Metacognition-know what you don’t know
- Actively seek knowledge
- Metalearning: learning algorithms that learn from other learning algorithms
- Identify general laws from specific cases
- Simulate and compare alternate virtual realities( models of the world)
The bigger question is why we need Artificial general intelligence at all, Is current narrow AI notenough. It turns out, AGI will solve the current limitations of narrow AI like bias, ethics, huge data, no generalization and allow us to solve problems more efficiently with less data. AGI/ Strong AI/Real AI will be one of the greatest strategic innovations in all human history just like electricity, computing and the internet. Beyond this AGI is a new space race, an exponential idea that will inspire the whole nation. Pursuing AGI will become India’s sputnik moment and place India in the global AI race along with China and USA. India needs to develop a comprehensive, high-priority plan for artificial general intelligence.
This will be a long continuous series to explore AGI and in our next articles we will be discussing more details
- How to reach from system 1 to system 2 deep learning?
- Capability maturity model of general intelligence.
- How companies like OpenAI and Deepmind are approaching AGI?
- Why does AI need a body?
- How will general intellice will evolve from a metaverse?
- General intelligence and theory of mind
- Tests for strong AI
- Consciousness Itself is Also an Unconscious
- Micro-PSI architecture
- Is Reinforcement learning enough for general intelligence
Madhav Bissa
Supriya Samuel
Raj Shekhar
Tarun Kumar
