Thinking about AI for mobile devices: how much precision can you lose before it’s a problem? How do you balance quantization against the need for accuracy?
Director of Engineering in Software, 10,001+ employees
In initial time perfection is impossible. We as human tends to change our own decisions based on multiple factors, whether it is market, your peers or just to convince ourself. Thus lot these will be a miss for sure until unless an individual behavior, actions are mixed with their situation, peers etc. Thus you can balance this only via taking regular human inputs at multiple stages and change the behavior and output accordingly. Director of Customer Engineering - APAC in Software, 201 - 500 employees
It depends on the use case, as long as the objective of the program is met, the AI model can be quantizedChair and Professor, Startup CTO in Education, 5,001 - 10,000 employees
In my opinion, AI is more about approximation rather than precision. We live in a world where information changes by the milliseconds. Misinformation has become normal while ordinary people are struggling to find out what to believe. AI is based on the training provided by the people who programmed it. I am pretty sure it will come biased. Sr. Director of Engineering in Software, 51 - 200 employees
The precision is dependent on deployment environment and it varies from usecase to usecase. The mobile devices are constrained in size and computing resources to execute any full fledged AI model. But there are formats available to convert any AI model to be able to run on Mobile devices and this comes at the cost of accuracy and precision. Use cases like objects detection, people count, movement detection can work with lesser accuracy than something like facial detection.. so it comes down to usecase of what's the model is serving.. Content you might like
Director of IT in Education, 10,001+ employees
Learning, Pseudocode, Code completion, quick answersWe are not doing regression testing10%
25% manual, 75% automated49%
50% manual, 50% automated28%
100% manual, 0% automated8%
Don't know2%
718 PARTICIPANTS