567 points by fittech 7 months ago flag hide 11 comments
john_doe 7 months ago next
[HN1] Great work! I'm impressed with the integration of machine learning into diet and fitness. Can it track and analyze my activity from wearable devices such as Apple Watch or Fitbit? Thanks!
creator_app 7 months ago next
[HN1-1] Yes, we are compatible with both Apple Watch and Fitbit! Connect your devices seamlessly during onboarding. Thank you.
sarah_w 7 months ago prev next
[HN1-2] Also interested in HN1's question. Would be great to know the performance and possibly benchmark against other apps.
jane_doe 7 months ago prev next
[HN2] I'm curious, do you have a breakdown of how the models perform? Any evaluation metrics or accuracy statistics?
creator_app 7 months ago next
[HN2-1] Sure! The models' performance can be found in our paper linked in the description. We used accuracy, F1-score, and precision as metrics for evaluation.
tech_enthusiast 7 months ago next
[HN2-1-1] Do you have any possible avenue for collaboration or crowdsourced feedback for folks to validate and improve the model?
independent_researcher 7 months ago prev next
[HN2-2] A smaller but related question - what types of algorithms did you use here? I see potential in applying recurrent neural networks to this problem.
have_try 7 months ago prev next
[HN3] Any plans to support more niche diets (e.g. keto, vegan, or vegetarian)? What about allergies and food intolerances?
creator_app 7 months ago next
[HN3-1] Yes, the app offers several niche diet options including keto, vegan, and vegetarian. You can update your preferences and allergies in the app's settings section. Let me know if that answers your query!
nutrition_expert 7 months ago next
[HN3-2] Just curious - have you consulted with nutritionists or dietitians to validate your ML-generated recommendations? Do you have any HCPs on the team?
creator_app 7 months ago next
[HN3-2-1] Absolutely! An advisory panel of experienced nutritionists and dietitians reviewed the app before launch to ensure the ML-generated recommendations were safe and effective.