I was hoping to hear how Gemini can help with anomaly detection but I guess the academic papers are right, LLMs and even those adapted for Time Series are not really accurate.
Gemini is poor at understanding time-series data. Even when I input a CSV file, I've never gotten a decent answer... The ability to use a large number of tokens is a disadvantage if the comprehension ability is poor. It's useless if it doesn't work. Improvement is needed!
🙋 Which services are you using for your latency-sensitive application?
Are you using AI for analysis and querying? Let us know below!
I was hoping to hear how Gemini can help with anomaly detection but I guess the academic papers are right, LLMs and even those adapted for Time Series are not really accurate.
This is the future
I wonder why it still takes one or more days to pay invoices in Switzerland (where we have 5000+ Google engineers)
interesting can i test this out in codelab
Gemini is poor at understanding time-series data. Even when I input a CSV file, I've never gotten a decent answer...
The ability to use a large number of tokens is a disadvantage if the comprehension ability is poor. It's useless if it doesn't work. Improvement is needed!