Hello DeepLearning Enthusiasts - looking for collaborators and learning partners

[jmyoung0711]

I’m Jeff and I’m thrilled to be part of this vibrant DeepLearning community. With a background rooted in finance, accounting, and a passion for strategy, business development, and negotiation, I bring a unique blend of skills to the AI and tech space.

My Background:

•	Graduated with honors and multiple scholarships in HBCOMM from Laurentian University.
•	Early career at KPMG
•	Executive roles such as CEO, CFO, COO, and VP Finance & Business Development.
•	Proven track record in change management, strategic planning, financial management, and board interactions.
•	Extensive experience with structured data, including building data scientist teams and developing models for actionable insights.

My Interest in AI:
While my expertise lies in finance and strategy, I have a keen interest in AI, data analytics, and their transformative potential across industries. I firmly believe in the power of AI to drive innovation, optimize processes, and unlock new opportunities. I am been using modeling and computer automation techniques long before AI was the norm. My expertise lies in structured data, systems and analytics.

Collaboration Opportunity:
I’m looking to collaborate with like-minded individuals passionate about AI and DeepLearning. Whether you’re a seasoned AI professional or a budding enthusiast, if you share a vision for leveraging AI to create impactful solutions, let’s connect! I’m open to discussing project ideas, exploring synergies, and working together to bring innovative AI projects to life.

How to Reach Me:
Feel free to send me a DM or email me at jmyoung0711@gmail.com Let’s collaborate, learn from each other, and make a positive impact through AI-powered solutions!

Looking forward to connecting with fellow AI enthusiasts and building something amazing together! #AICommunity #CollaborateInAI #DeepLearning

[Nevermnd]

Hey Jeff,

Welcome.

My original (misguided) ‘formal’ background is in Finance/Economics as well (I spent too much time trying to ‘not be a nerd’, before finally accepting, well that is what I am after all. Can’t escape fate.)

I am somewhat newer to the ML/‘AI’ side, at least as it is generally taught today (i.e. Deep Networks versus, then, decision trees), but have a long background in programming (C/C++, Python, Javascript, R, SQL) and hardware (design and Verilog/Systems Verilog).

As much as I find this whole topic really interesting and exciting, I would, however say perhaps I am a little in the ‘conservative’ camp (?).

I mean I certainly don’t see Deep Learning as the ‘sledge hammer for all problems’, and find the near term chance of AGI highly dubious (perhaps a ‘simulacrum’, but then that’s not really ‘AGI’ now is it ?). Or lets just say I am wary of the hype cycle.

In any case, again, welcome !

[jmyoung0711]

Pleasure to meet you. I am a believer in AI and Machine Learning not because of the technology hype but because of the principles. This might sound odd but in my view AI is very much based on the principles of Lean or Six Sigma which I have been using for years. You have an input a process and an output. The goal is to be as efficient and accurate as possible. This is DMAIC at its core.

Although having a finance background I primarily became operations and systems focused. Ironically in industries where I had little or no background. A recent example was in the agricultural space. We had challenges with yield and quality but apparently also an industry guru at the helm. I read and researched and found some interesting data points and still resistance. So I brought in a data team who analyzed the vast amounts of data we had from temperature, to humidity, to water usage etc… they found valuable insights… adjustments were made and yields improved…

So for me the two biggest things that AI and Machine Learning can do is increase efficiency and eliminate human bias. We shouldn’t be afraid of these tools. They provide us a way and a method to be better than we ever could have in the past. They don’t replace people they improve people. AI doesn’t work without technology, business and subject matter experts working together. For some reason it is being feared but it shouldn’t be.

I can code just enough to be dangerous, I have an excellent understanding of structured data, unstructured I understand but how to manipulate it still learning. Lean principles very high knowledge… passion to work with some like minded people to tackle projects is high… what they are not sure… but why take the journey alone.

Just my opinion from the cheap seats.

[TMosh]

@jmyoung0711

for me the two biggest things that AI and Machine Learning can do is increase efficiency and eliminate human bias.

One observation and caution:

A significant risk in AI is that a “human bias” is replaced by “dataset bias”.

A model can only learn from the data it is presented. If the dataset has built-in biases (either in how the dataset was collected, or how the dataset was labeled), then this will influence its predictions.

[markiedoo]

I’m new to Ai learning I want to learn more from you Jeff

[Nevermnd]

@jmyoung0711 No problem Jeff, and your opinions/input/experience/questions are more than welcome here.

In my mind though, while ‘efficiency’ can be a component or end goal of Deep Nets, I am not sure that is exactly that is their stated purpose-- Or even, really, precisely what they excel at.

Rather I see their innate strength lying in ‘pattern matching’.

I mean I recall earlier on in my Data Science studies, as an exercise, having to work on a recommender system for the MovieLens 10M data set. Now this data set is just so huge (yet even still small by today’s ‘Big Data’ standards) to the extent that to produce useful insights…

Well, for one there is no way you can just pour over the data manually as you might with a much smaller set, it’d take you years. And even then you probably would not gain any useful intuition. Second, it is also both to large and the relationships too complex to graph or ‘visualize’ in a way which will get much further either.

Yet, with the right techniques you can garner patterns in the data, seek insights, and then rigorously test them. Or I tell others that for me learning Data Science was like ‘learning to see when you are blind’.

Further, as a set of tools it allows us to probe for information on problems that are simply too complex for any one person to possibly solve.

Another problem I worked on was Malware detection, even on a small data set (n ~= 5000)-- but with 150 features (independent variables). End detection was at a 99% accuracy level, but ‘by hand’, how could a person possibly parse that in their head ?

@TMosh, however, does bring up a really, really important point about the tradeoff bias. Increasingly ‘ethicists’ (and my first degree was in Philosophy, so I feel I am allowed a little snipe here) in AI talk about the ‘alignment’ problem.

I mean, yes, whom and how the model is constructed, particularly in traditional ML can be a serious issue-- But even in DNNs it is kind of like ‘You don’t have an ‘alignment problem’, you have a dataset bias problem’.

For some concrete examples of this, last year I read Cathy O’Neil’s ‘Weapons of Math Destruction’. Already aware of some of these issues I still thought it was a good read.

Though, while not quite ‘removing’ human bias, one could say this aptitude at pattern matching does, potentially, conceptualize in ways we usually don’t (i.e. I think of DeepMind’s win at the game of Go).

So, as to ‘efficiency’… --sometimes; And I actually have more thoughts on a different side/aspect of this issue yet will save for later.

In the meantime, I, at least, do not want to become a paperclip.