I work for an event management platform and I am working on a recommendation system based on Machine Learning. I'm really enjoying working with this stuff although sometimes it can get pretty complex. Anyone else here working with this stuff?
@rsk said:
Im working on something similar to target customers with personalized push notifications based on historical data. Not hosting related.
What tools are you using?
At the moment I am testing two approaches: one using a Ruby library for recommendations, so I only need to prepare the training data correctly, and the other is with GCP's services for ML.
The recommendations in our case are a bit more complex. We need to recommend other attendees in an event by matching interests, but the weirdness/complication is that we want to rank higher opposite intents compared to other matches. So for example if I am looking for an investment in X, I want to rank higher who is offering an investment in X. So it works differently from usual recommendation engines.
How much of the complexity is code and how much is advanced math?
In my case most of the complexity is in producing the training data due to our particular stuff. For the math part like I said I just use a library that handles that for me. I don't know math well enough yet but I am planning to learn because I want to build a new recommendations library for Crystal.
@drizbo said:
Today its really modern to call every 100 lines of code artificial intelligence.
@vitobotta said: Are you happy with the course? Do you recommend it?
There are lot of positive reviews but I felt like I'm missing something. It's good for introductory theory so I'm pushing through. Perhaps I will try building something first before getting back to theory after finishing this course.
@vitobotta said: Are you happy with the course? Do you recommend it?
There are lot of positive reviews but I felt like I'm missing something. It's good for introductory theory so I'm pushing through. Perhaps I will try building something first before getting back to theory after finishing this course.
@vitobotta said:
I work for an event management platform and I am working on a recommendation system based on Machine Learning. I'm really enjoying working with this stuff although sometimes it can get pretty complex. Anyone else here working with this stuff?
I have experience working with machine learning-based trading systems in financial markets. This field can be challenging due to the complexities of market data and the need for accurate predictions, but it can also be highly rewarding to see your models generate profitable trades.
To succeed in this field, it's important to continually seek out new research and best practices and to be open to trying out different approaches until you find what works best for your specific use case. It's also essential to stay up-to-date with the latest advancements in the field, as machine learning is constantly evolving and improving.
Overall, working with machine learning can be an exciting and intellectually stimulating experience. I'm glad to hear that you're enjoying it too!
Comments
Im working on something similar to target customers with personalized push notifications based on historical data. Not hosting related.
What tools are you using?
At the moment I am testing two approaches: one using a Ruby library for recommendations, so I only need to prepare the training data correctly, and the other is with GCP's services for ML.
The recommendations in our case are a bit more complex. We need to recommend other attendees in an event by matching interests, but the weirdness/complication is that we want to rank higher opposite intents compared to other matches. So for example if I am looking for an investment in X, I want to rank higher who is offering an investment in X. So it works differently from usual recommendation engines.
I'm learning. Almost completed Supervised Machine Learning course by Andrew Ng and taking NLP intro next.
How much of the complexity is code and how much is advanced math?
Today its really modern to call every 100 lines of code artificial intelligence.
Are you happy with the course? Do you recommend it?
In my case most of the complexity is in producing the training data due to our particular stuff. For the math part like I said I just use a library that handles that for me. I don't know math well enough yet but I am planning to learn because I want to build a new recommendations library for Crystal.
What do you mean?
There are lot of positive reviews but I felt like I'm missing something. It's good for introductory theory so I'm pushing through. Perhaps I will try building something first before getting back to theory after finishing this course.
Does it cover the math?
Yea, but on a surface level. It's a introductory course so don't expect much.
Ok. I am looking for a course or something that explains the math in a "for dummies" format
I have experience working with machine learning-based trading systems in financial markets. This field can be challenging due to the complexities of market data and the need for accurate predictions, but it can also be highly rewarding to see your models generate profitable trades.
To succeed in this field, it's important to continually seek out new research and best practices and to be open to trying out different approaches until you find what works best for your specific use case. It's also essential to stay up-to-date with the latest advancements in the field, as machine learning is constantly evolving and improving.
Overall, working with machine learning can be an exciting and intellectually stimulating experience. I'm glad to hear that you're enjoying it too!
It's common to label something that generates content or makes a decision as 'AI', when in fact it's just a procedural script/app.
This sounds annoying af. Hopefully can opt-out?