Jl Hills

Invited review: An evaluation of the likely effects of individualized feeding of concentrate supplements to pasture-based dairy cows

In pasture-based dairy systems, supplementary feeds are used to increase dry matter intake and milk production. Historically, supplementation involved the provision of the same amount of feed (usually a grain-based concentrate feed) to each cow in the herd during milking (i.e., flat-rate feeding). The increasing availability of computerized feeding and milk monitoring technology in milking parlors, however, has led to increased interest in the potential benefits of feeding individual cows (i.e., individualized or differential feeding) different amounts and types of supplements according to one or more parameters (e.g., breeding value for milk yield, current milk yield, days in milk, body condition score, reproduction status, parity). In this review, we consider the likely benefits of individualized supplementary feeding strategies for pasture-based dairy cows fed supplements in the bail during milking. A unique feature of our review compared with earlier publications is the focus on individualized feeding strategies under practical grazing management. Previous reviews focused primarily on research undertaken in situations where cows were offered ad libitum forage, whereas we consider the likely benefits of individualized supplementary feeding strategies under rotational grazing management, wherein pasture is often restricted to all or part of a herd. The review provides compelling evidence that between-cow differences in response to concentrate supplements support the concept of individualized supplementary feeding.

Dynamic cattle behavioural classification using supervised ensemble classifiers

Cattle behavioural classification using cattle tag and supervised ensemble classifiers.Unsupervised hybrid clustering used to study inherent natural grouping in data set.Best classification accuracy was 96% using the bagging ensemble with Tree learner.A mechanism for the early detection and quantitative assessment of animal health. In this paper various supervised machine learning techniques were applied to classify cattle behaviour patterns recorded using collar systems with 3-axis accelerometer and magnetometer, fitted to individual dairy cows to infer their physical behaviours. Cattle collar data was collected at the Tasmanian Institute of Agriculture (TIA) Dairy Research Facility in Tasmania. In the first stage of analysis a novel hybrid unsupervised clustering framework, comprised of probabilistic principal component analysis, Fuzzy C Means, and Self Organizing Map network algorithms was developed and used to study the natural structure of the sensor data. Findings from this unsupervised clustering were used to guide the next stage of supervised machine learning. Five major behaviour classes, namely, Grazing, Ruminating, Resting, Walking, and other behaviour were identified for the classification trials. An ensemble of classifiers approach was used to learn models of cow behaviour using sensor data and ground truth behaviour observations acquired from the field. Ensemble classification using bagging, Random Subspace and AdaBoost methods along with conventional supervised classification methods, namely, Binary Tree, Linear Discriminant Analysis classifier, Naive Bayes classifier, k-Nearest Neighbour classifier, and Adaptive Neuro Fuzzy Inference System classifier were compared. The highest average correct classification accuracy of 96% was achieved using the bagging ensemble classification with Tree learner, which had 97% sensitivity, 89% specificity, 89% F1 score and 9% false discovery rate. This study has shown that cattle behaviours can be classified with a high accuracy using supervised machine learning technique. As dairy and beef systems become more intensive, the ability to identify the changes in the behaviours of individual livestock becomes increasingly difficult. Accurate behavioural monitoring through sensors provides a significant potential in providing a mechanism for the early detection and quantitative assessment of animal health issues such a lameness, informing key management events such as the identification of oestrus, or informing changes in supplementary feeding requirements.

Bag of Class Posteriors, a new multivariate time series classifier applied to animal behaviour identification

A new multi-scale time series classifier is proposed using class posterior estimates.The classifier infers a large set of animal behaviour using motion based time series.The proposed classifier outperforms benchmark classifiers by between 43% and 77%.The proposed classifier is found to be more efficient than the Bag of Features model. In this paper, two new multivariate time series classifiers are introduced as the Bag of Class Posteriors (BOCP) and the Bag of Class Posterior with Ordering (BOCPO). The models propose a new multi-scale feature representation where the class posterior estimates of contiguous local patterns are aggregated over longer time scales. The models are employed as part of an animal behaviour monitoring system that are comprised of sensors, which are fitted to the animals, and a classifier that translates sensor data into knowledge of the animals behaviour.Animal monitoring systems are commonly developed to infer a small number of behaviours with relevance to a specific application. To investigate if a standard monitoring system with an Inertial Measurement Unit (IMU) can be reused for different management applications, a set of ten cattle behaviours relevant to different management applications were classified with the proposed models. Results indicate that the multi-scale BOCP and BOCPO models were far more capable of classifying the cow behaviours offering a 43% to 77% improvement over benchmark time interval classifiers with fixed time resolution. In addition, the BOCPO model was shown to offer a far more efficient feature representation than the related multi-scale Bag of Features (BOF) classifier (up to 200 times smaller) making it better suited to deploy upon monitoring devices fitted to animals in the field.

Conditioned feeding responses of sheep towards flavoured foods associated with casein administration: the role of long delay learning

The objective of two experiments was to investigate whether a delayed type of learning could account for the conditioned feeding responses of sheep towards novel food flavours associated with post-ingestive consequences (PIC) created from the administration at different points in time of a nutritive stimulus (casein). The doses of casein were low (15 g) and high (75 g) for experiments 1 and 2, previously known to result in positive and negative PIC respectively. Each experiment consisted of three conditioning periods, during which sheep were trained to associate one of two novel flavours with either casein or water (placebo) administration. During each conditioning, a novel flavoured food low in protein and relatively high in energy, was offered for 3 h (08:00 to 11:00 h) and was followed by an unflavoured, nutritionally similar food for the rest of the feeding time (11:00 to 17:00 h). Sheep were randomly assigned to one of three treatments that were defined by the time when casein or water doses were administered, in relation to the presence of the flavoured food (A= 08:30 and 10:00, B= 11:30 and 13:00 and С = 14:30 and 16:00 h respectively). At the end of each conditioning period preference tests were performed, where sheep were offered a choice between the two flavoured foods. There was no effect of time of casein administration on the conditioned responses towards flavoured foods in either experiment. In both experiments, the proportion of the flavoured food selected was significantly affected by the interaction between preference tests and casein association. For experiment 1 this was due to an increasing preference for the casein associated food accompanied by a decreasing preference for the water associated flavoured food as a result of repeated conditioning. The degree of such preference was different between flavours used for association with casein or water. For experiment 2 avoidance of the casein and preference for the water associated food were established after the completion of the second and reinforced by the third conditioning period. Flavours used had a lesser effect on the conditioned responses of this experiment. The results support the view that sheep develop conditioned responses towards novel food flavours associated with the administration of a nutritive stimulus, even when the PIC resulting from its administration are significantly disassociated in time from the presence of the flavoured food.

Conditioned feeding responses in sheep to flavoured foods associated with sulphur doses.

A study was conducted to determine whether sheep form conditioned flavour aversions (CFAs) or preferences (CFPs) for food flavours associated, respectively, with excessive or appropriate concentrations of sulphur (S) and also whether the rate of formation and strength of CFAs and CFPs are dependent on the animals initial S status or the level of administration of S. In experiment 1, 48 mature ewes were conditioned to associate a new food containing a novel flavour with an infusion of S delivered intra-ruminally, or the same food containing another novel flavour with an infusion of distilled water. The same flavours were then used in experiment 2. At the end of each conditioning period, the relative preference for the two flavoured foods was determined by measuring the amount of each food ingested during a two-choice, 20-min preference test. Experiment 1 consisted of two phases. In phase 1 each conditioning period lasted for 5 days and was repeated four times, whereas in phase 2 the conditioning period lasted for 8 days and was repeated three times. In experiment 1 the sheep were initially in an S-adequate state. In experiment 2, the sheep were re-randomized to treatments and started in an S-depleted state. The conditioning periods also lasted for 8 days and were repeated three times. There was no evidence to support the hypothesis that sheep develop CFAs or CFPs to food flavours associated with S doses in phase 1 of experiment 1. In phase 2, however, sheep formed CFAs towards the food with the flavour they had come to associate with administration of high levels of S. Repeated exposure to the flavour associated with high levels of S led to stronger aversions and there was an interaction between the S dose level and conditioning periods, indicating that the rate of development of these CFAs was highest for the highest S dose levels. The differences between results of phase 1 and 2 were probably due to the different numbers of reinforcements and different intervals between specific flavour/dose associations. In experiment 2 there was no evidence for the development of CFPs or CFAs to food flavours associated with S doses. The apparent indifference of the sheep to S was probably due to their responding more to their previous experience of the food flavours than to their S status. Spearman rank correlations on flavour preferences indicated that conditioned flavour responses formed in experiment 1 persisted in individual sheep when they were allocated at random into their new treatments in experiment 2 and influenced or masked the formation of new associations. This demonstration of ‘carry-over’ effects highlights the importance of considering an animal’s previous experience of flavours and their associations with post-ingestive consequences when coming to conclusions concerning current development of CFAs and CFPs. These results may also have more general implications for feeding studies in animals that are randomized into treatment groups without regard to their previous feeding experiences.

More milk from forage: Milk production, blood metabolites, and forage intake of dairy cows grazing pasture mixtures and spatially adjacent monocultures

There is interest in the reincorporation of legumes and forbs into pasture-based dairy production systems as a means of increasing milk production through addressing the nutritive value limitations of grass pastures. The experiments reported in this paper were undertaken to evaluate milk production, blood metabolite concentrations, and forage intake levels of cows grazing either pasture mixtures or spatially adjacent monocultures containing perennial ryegrass (Lolium perenne), white clover (Trifolium repens), and plantain (Plantago lanceolata) compared with cows grazing monocultures of perennial ryegrass. Four replicate herds, each containing 4 spring-calving, cross-bred dairy cows, grazed 4 different forage treatments over the periods of early, mid, and late lactation. Forage treatments were perennial ryegrass monoculture (PRG), a mixture of white clover and plantain (CPM), a mixture of perennial ryegrass, white clover, and plantain (RCPM), and spatially adjacent monocultures (SAM) of perennial ryegrass, white clover, and plantain. Milk volume, milk composition, blood fatty acids, blood β-hydroxybutyrate, blood urea N concentrations, live weight change, and estimated forage intake were monitored over a 5-d response period occurring after acclimation to each of the forage treatments. The acclimation period for the early, mid, and late lactation experiments were 13, 13, and 10 d, respectively. Milk yield (volume and milk protein) increased for cows grazing the RCPM and SAM in the early lactation experiment compared with cows grazing the PRG, whereas in the mid lactation experiment, milk fat increased for the cows grazing the RCPM and SAM when compared with the PRG treatments. Improvements in milk production from grazing the RCPM and SAM treatments are attributed to improved nutritive value (particularly lower neutral detergent fiber concentrations) and a potential increase in forage intake. Pasture mixtures or SAM containing plantain and white clover could be a strategy for alleviating the nutritive limitations of perennial ryegrass monocultures, leading to an increase in milk production for spring calving dairy cows during early and mid lactation.

Detecting heat events in dairy cows using accelerometers and unsupervised learning

We developed a heat detection algorithm for pasture-based dairy cows.Our algorithm uses data from accelerometer attached to the cow collars.We present the overall accuracy of 82-100% with 100% sensitivity. This study was conducted to investigate the detection of heat events in pasture-based dairy cows fitted with on-animal sensors using unsupervised learning. Accelerometer data from the cow collars were used to identify increased activity levels in cows associated with recorded heat events. Time series data from the accelerometers were first segmented into windows before features were extracted. K-means clustering algorithm was then applied across the windows for grouping. The groups were labelled in terms of their activity intensity: high, medium and low. An activity index level (AIxL) was then derived from a count of activity intensity labels over time. Change detection techniques were then applied on AIxL to find very high activity events. Detected events in AIxL were compared with recorded heat events and observed significant associations between the increased activities through high AIxL values and the observed heat events. We achieved overall accuracy of 82-100% with 100% sensitivity when change detection technique is applied to activity index level.

Behavior classification of cows fitted with motion collars: Decomposing multi-class classification into a set of binary problems

Precision management systems for livestock offer the potential to monitor and manage animals on an individual basis. A key component of these sensor based systems are the analytical models that automatically translate sensor data into different behavioral categories. A new methodology was proposed for multi-class behavior modeling based upon the “one-vs-all” framework. This methodology differs from the standard approach to behavior classification where a single classifier is trained to discriminate between multiple behaviors. Instead a set of binary classifiers are trained to each discriminate one of the behavior classes against a combined class of all the remaining behaviors. The confidence scores from the set of binary classifiers are then combined to generate a behavioral estimate. The performance of this new modeling approach is validated across a study involving 24 Holstein-Friesian dairy cows that were each fitted with an Inertial Measurement Unit (IMU) on a collar upon their neck. Five general classes of cow behavior grazing, walking, ruminating, resting and “Other” were classified. Binary time series classifiers were tailored to each of the five behaviors through training and validation of each model across a range of configurations including nine window sizes, five classifiers and a feature selection process of 84 candidate input features. Results revealed that the proposed model classified grazing behavior with an extremely high classification accuracy (F-score of 0.98), whilst ruminating and resting behaviors were also classified with a high accuracy (F-scores of 0.87 and 0.85, respectively), and walking was classified with a lower accuracy (F-score of 0.73). The proposed model offered a 5% performance improvement over standard multi-class time series classifiers that was attributed to the “one-vs-all” framework training a classifier for each behavior independently; this created diversity in the behavior model. The feature selection process used in developing each of the binary classifiers found that features representing the motion intensity and pitch of the cow’s head were most important to each of behavior’s classification. Whilst a minor performance improvement was obtained using the proposed methodology, it is suggested that further performance improvements could be obtained by increasing the diversity of the classifier’s inputs. Diversity could be created by fusing the data of other sensors that can be fitted to the cows i.e. GPS tracking unit, pressure sensor and microphone.

A comparison of autoencoder and statistical features for cattle behaviour classification.

A study is presented comparing the effectiveness of unsupervised feature representations with handcrafted features for cattle behaviour classification. Precision management of cattle requires the interaction of individual animals to be continuously monitored on the farm. Consequently, classifiers are trained to infer the behaviour of the animals using the observations from the sensors that are fitted upon them. Historically, domain knowledge drives the generation of features for cattle behaviour classifiers. When new behaviours are introduced into the system, however, it is often necessary to modify the feature set; this requires additional design and more data. Autoencoders, on the other hand, can skip this design step by learning a common, unsupervised feature representation for training. Whilst stacked autoencoders successfully represent structured data including speech, language and images, deep networks have not been used to model cattle motion. Hence, we investigate using a stacked autoencoder to learn a feature representation for cattle behaviour classification. Experimental results demonstrate that the autoencoder features perform reasonably well in comparison to the statistical features that are selected using prior knowledge of behaviour motion.

An investigation of cow feeding behavior using motion sensors

An experiment to study the impact of supplements upon the feeding behavior of dairy cattle was conducted at the Tasmanian Institute of Agriculture (TIA) Dairy Research Facility. Collar systems with 3-axis accelerometer and magnetometer were fitted to individual cows to infer their feeding behavior. We describe the solutions applied to correct for sensor data issues, and then provide some preliminary analysis associated with developing behavior models using multivariate time series data.