A Gaussian process regression model to predict energy contents of corn for poultry

Application of Machine Learning Algorithms to Describe the Characteristics of Dairy Sheep Lactation Curves

In recent years, machine learning (ML) algorithms have emerged as powerful tools for predicting and modeling complex data. Therefore, the aim of this study was to evaluate the prediction ability of different ML algorithms and a traditional empirical model to estimate the parameters of lactation curves. A total of 1186 monthly records from 156 sheep lactations were used. The model development process involved training and testing models using ML algorithms. In addition to these algorithms, lactation curves were also fitted using the Wood model. The goodness of fit was assessed using correlation coefficient (r), mean absolute error (MAE), root mean square error (RMSE), relative absolute error (RAE), and relative root mean square error (RRSE). SMOreg was the algorithm with the best estimates of the characteristics of the sheep lactation curve, with higher values of r compared to the Wood model (0.96 vs. 0.68) for the total milk yield. The results of the current study showed that ML algorithms are able to adequately predict the characteristics of the lactation curve, using a relatively small number of input data. Some ML algorithms provide an interpretable architecture, which is useful for decision-making at the farm level to maximize the use of available information.

Resazurin-based high-throughput screening method for the discovery of dietary phytochemicals to target microbial transformation of L-carnitine into trimethylamine, a gut metabolite associated with cardiovascular disease

Nowadays, there is great interest in the discovery of food compounds that might inhibit gut microbial TMA production from its methylamine precursors. In this work, an innovative novel screening strategy capable of rapidly determining the differences in the metabolic response of Klebsiella pneumoniae, a bacteria producing TMA under aerobic conditions, to a library of extracts obtained from food and natural sources was developed. The proposed high-throughput screening (HTS) method combines resazurin reduction assay in 384-well plates and Gaussian Processes as a machine learning tool for data processing, allowing for a fast, cheap and highly standardized evaluation of any interfering effect of a given compound or extract on the microbial metabolism sustained by L-carnitine utilization. As a proof-of-concept of this strategy, a pilot screening of 39 extracts and 6 pure compounds was performed to search for potential candidates that could inhibit in vitro TMA formation from L-carnitine. Among all the extracts tested, three of them were selected as candidates to interfere with TMA formation. Subsequent in vitro assays confirmed the potential of oregano and red thyme hexane extracts (at 1 mg mL^-1) to inhibit TMA formation in bacterial lysates. In such in vitro assay, the red thyme extract exerted comparable effects on TMA reduction (∼40%) as 7.5 mM meldonium (∼50% TMA decrease), a reported L-carnitine analogue. Our results show that metabolic activity could be used as a proxy of the capacity to produce TMA under controlled culture conditions using L-carnitine to sustain metabolism.

Comparison of Machine Learning Algorithms for Structure State Prediction in Operational Load Monitoring

The aim of operational load monitoring is to make predictions about the remaining usability time of structures, which is extremely useful in aerospace industry where in-service life of aircraft structural components can be maximized, taking into account safety. In order to make such predictions, strain sensors are mounted to the structure, from which data are acquired during operational time. This allows to determine how many load cycles has the structure withstood so far. Continuous monitoring of the strain distribution of the whole structure can be complicated due to vicissitude nature of the loads. Sensors should be mounted in places where stress and strain accumulations occur, and due to experiencing variable loads, the number of required sensors may be high. In this work, different machine learning and artificial intelligence algorithms are implemented to predict the current safety factor of the structure in its most stressed point, based on relatively low number of strain measurements. Adaptive neuro-fuzzy inference systems (ANFIS), support-vector machines (SVM) and Gaussian processes for machine learning (GPML) are trained with simulation data, and their effectiveness is measured using data obtained from experiments. The proposed methods are compared to the earlier work where artificial neural networks (ANN) were proven to be efficiently used for reduction of the number of sensors in operational load monitoring processes. A numerical comparison of accuracy and computational time (taking into account possible real-time applications) between all considered methods is provided.