A survey on computational taste predictors.

Area of Science:

• Sensory science
• Computational chemistry
• Food science

Background:

• Taste perception is vital for distinguishing safe foods from toxins, involving five basic tastes: sweet, bitter, umami, salty, and sour.
• Predicting taste is complex due to the multidisciplinary nature of taste perception and numerous influencing factors.
• Machine learning (ML) offers powerful tools to analyze complex taste data and uncover hidden correlations.

Purpose of the Study:

• To review recent advancements in taste prediction methodologies.
• To analyze existing food-related databases relevant to taste prediction.
• To survey the latest taste prediction tools developed in recent years.

Main Methods:

• Literature review of scientific publications on taste prediction.
• Analysis of publicly available food composition and taste databases.
• Evaluation of machine learning algorithms applied to taste prediction.
• Review of existing computational tools for taste prediction.

Main Results:

• Machine learning has significantly improved the ability to predict taste by integrating large datasets and identifying complex patterns.
• Several food-related databases and taste prediction tools have emerged, offering valuable resources for research and industry.
• The review highlights the growing integration of computational approaches in understanding taste perception.

Conclusions:

• ML-driven taste prediction is a rapidly advancing field with significant implications for various industries.
• Further development of databases and predictive models is crucial for enhancing taste prediction accuracy.
• This review provides a comprehensive overview of the current state and future directions in taste prediction research.