James Hensman

We introduce stochastic variational inference for Gaussian process models. This enables the application of Gaussian process (GP) models to data sets containing millions of data points. We show how GPs can be vari- ationally decomposed to depend on a set of globally relevant inducing variables which factorize the model in the necessary manner to perform variational inference. Our ap- proach is readily extended to models with non-Gaussian likelihoods and latent variable models based around Gaussian processes. We demonstrate the approach on a simple toy problem and two real world data sets.

Gaussian process classification is a popular method with a number of appealing properties. We show how to scale the model within a variational inducing point framework, outperforming the state of the art on benchmark datasets. Importantly, the variational formulation can be exploited to allow classification in problems with millions of data points, as we demonstrate in experiments.

GPflow is a Gaussian process library that uses TensorFlow for its core computations and Python for its front end. The distinguishing features of GPflow are that it uses variational inference as the primary approximation method, provides concise code through the use of automatic differentiation, has been engineered with a particular emphasis on software testing and is able to exploit GPU hardware.

OBJECTIVE: To characterize the circadian clock in murine cartilage tissue and identify tissue-specific clock target genes, and to investigate whether the circadian clock changes during aging or during cartilage degeneration using an experimental mouse model of osteoarthritis (OA). METHODS: Cartilage explants were obtained from aged and young adult mice after transduction with the circadian clock fusion protein reporter PER2::luc, and real-time bioluminescence recordings were used to characterize the properties of the clock. Time-series microarrays were performed on mouse cartilage tissue to identify genes expressed in a circadian manner. Rhythmic genes were confirmed by quantitative reverse transcription-polymerase chain reaction using mouse tissue, primary chondrocytes, and a human chondrocyte cell line. Experimental OA was induced in mice by destabilization of the medial meniscus (DMM), and articular cartilage samples were microdissected and subjected to microarray analysis. RESULTS: Mouse cartilage tissue and a human chondrocyte cell line were found to contain intrinsic molecular circadian clocks. The cartilage clock could be reset by temperature signals, while the circadian period was temperature compensated. PER2::luc bioluminescence demonstrated that circadian oscillations were significantly lower in amplitude in cartilage from aged mice. Time-series microarray analyses of the mouse tissue identified the first circadian transcriptome in cartilage, revealing that 615 genes (∼3.9% of the expressed genes) displayed a circadian pattern of expression. This included genes involved in cartilage homeostasis and survival, as well as genes with potential importance in the pathogenesis of OA. Several clock genes were disrupted in the early stages of cartilage degeneration in the DMM mouse model of OA. CONCLUSION: These results reveal an autonomous circadian clock in chondrocytes that can be implicated in key aspects of cartilage biology and pathology. Consequently, circadian disruption (e.g., during aging) may compromise tissue homeostasis and increase susceptibility to joint damage or disease.