In the not-so-distant past, machines were relegated to mere calculation and automation. Today machines can create, innovate, and even rival human imagination! Is GenAI slowly dissolving the lines between human creativity and artificial intelligence? Let's journey through the early days of Generative AI...
Before that, here is a little back story of "how I met Bangaly"
I have known and admired Bangaly's professional accomplishments for more than a year now. He has over 1000 Badges, a CNCF Kubestronaut, a AWS Golden Jacket , 17xAWS, 14xGCP, 17xMicrosoft, 6xComptia, 4xHashicorp, 4xGithub, 3xOCI, and many more. I couldn't comprehend how he achieved that, so I did the only logical thing, ask him!
The answer wasn't anything we all haven't heard before. There was the irritating consistency, the annoying hardwork, never missing curiosity and a the magical kindness
View the crazy things he has accomplished here ☞Bangaly's LinkedIn
We are kind of an odd duo, with a shared passion to encourage, empower and share knowledge with our community. We shall be creating series of content surrounding the realm of AI, ML and all the fun stuff around Cloud Computing.
Let's start from the beginning.....
Artificial Intelligence has been lingering around since 1941!
In 1950, Alan Turing published a paper titled "Computing Machinery and Intelligence" in which he proposed the imitation game.
The game involves: Three participants: a human interrogator, a human respondent, and a machine.
The interrogator's goal is to determine which of the other two participants is the human and which is the machine.
Communication is limited to text-only exchanges.
The machine aims to fool the interrogator into thinking it is human. Turing argued that if the machine could consistently fool human interrogators, it should be considered to exhibit intelligent behavior. The term Artificial Intelligence was officially coined on august 31, 1955. Here is the link to the article
What then is machine learning?
Machine Learning is a subset of Artificial intelligence. The idea behind machine learning is to feed large amounts of data into algorithms, which can then identify patterns and relationships in the data, and use that knowledge to make predictions or decisions without relying on hard-coded rules. Arthur Samuel, an IBM researcher, is credited with coining the term "Machine Learning" in 1959.
and Deep Learning is...
Deep Learning is a specific type of machine learning that gained widespread attention in the 21st century. It is inspired by the structure and function of the human brain and involves the use of artificial neural networks (ANNs), which are computational models that mimic the interconnected neurons in the brain.
Deep learning algorithms can automatically learn complex patterns and representations from raw data, making them highly effective for tasks like image recognition, natural language processing, and speech recognition. While the fundamental concepts of neural networks date back to the 1940s and 1950s, with pioneers like Warren McCulloch and Walter Pitts, deep learning only became practically viable in the late 2000s due to advances in computing power, availability of large datasets, and algorithmic improvements.
In summary,
AI is the overarching field, machine learning is a subset of AI that focuses on learning from data, and deep learning is a specific type of machine learning that uses artificial neural networks.
These fields have evolved over time, with each advancement building upon the foundations laid by previous researchers and breakthroughs.
A Question we should ask ourselves is, if AI has been around for this long, why is it gaining popularity decades after?
Early AI researchers faced several significant challenges as they worked to develop artificial intelligence in the 1950s and 1960s. These included:
Limited Computing Power: Early computers lacked the processing power and memory needed for complex AI algorithms and large data processing.
Unrealistic Expectations: Early AI pioneers were overly optimistic about achieving human-level AI quickly, leading to disappointment when progress was slower.
Lack of Understanding of Intelligence: Researchers underestimated the complexity of human cognition, thinking it could be easily replicated in machines.
Narrow Focus: Early AI research focused on specific tasks like chess, which didn’t translate well to general intelligence or common sense reasoning.
Funding Challenges: AI research experienced cycles of enthusiasm and funding, followed by “AI winters” where interest and financial support dried up.
Software Limitations: The bottleneck in AI development was often software, as creating programs that mimicked human reasoning was very difficult.
Lack of Data: Early researchers didn’t have access to the large datasets needed for effective AI training, limiting their models’ effectiveness.
How have we overcomed these challenges?
Access to Artificial Intelligence has been democratized and its capabilities are continuously advancing. It has also led to generative AI, a game-changing technology.
Generative AI refers to artificial intelligence models that can generate new data, such as texts, images, audios, or videos, based on the patterns and relationships learned from training data, and it is still evolving.
One of the key players in democratizing access to generative AI is AWS Bedrock, a comprehensive platform developed by Amazon Web Services (AWS) that provides developers and researchers with the tools, infrastructure, and resources needed to build and deploy generative AI models.
AWS Bedrock leverages the power of cloud computing, offering scalable and cost-effective solutions that enable users to train and run large-scale generative AI models without the need for expensive on-premises hardware. By removing barriers to entry, AWS Bedrock has empowered individuals and organizations of all sizes to explore and leverage the capabilities of generative AI.
AWS Bedrock use-case.
Let's illustrate one of the capability of AWS Bedrock using AWS management console.
How to create an Image on AWS Bedrock
This can be done in 3 simple steps
Step 1
You require an AWS Account, which is free. Be aware that most of the models are third party tools and will be charged separately, even though you have AWS Credits. We learned the hard way 🙃💸
Log into the account and move to the AWS Bedrock UI.
AWS Bedrock UI and its functionalities
Step 2
Request for access from the desired model
Step 3
Image generated from desired image
AWS Samples contains pre-built examples to help customers get started with the Amazon Bedrock service.
We shall be exploring more of them in our next content.
Stay tuned....
Bangaly and Endah
Top comments (1)
I met Bangaly at the Google Next 2024 conference and can confirm his 'magical kindness'. His positive energy is truly magnetic. It's wonderful to see someone who combines such technical excellence with genuine warmth and a passion for community building.
Looking forward to following this content series!