Adaptive systems -- such as a biological organism gaining survival advantage, an autonomous robot executing a functional task, or a motor protein transporting intracellular nutrients -- must model the regularities and stochasticity in their environments to take full advantage of thermodynamic resources. Analogously, but in a purely computational realm, machine learning algorithms estimate models to capture predictable structure and identify irrelevant noise in training data. This happens through optimization of performance metrics, such as model likelihood. If physically implemented, is there a sense in which computational models estimated through machine learning are physically preferred? We introduce the thermodynamic principle that work production is the most relevant performance metric for an adaptive physical agent and compare the results to the maximum-likelihood principle that guides machine learning. Within the class of physical agents that most efficiently harvest energy from their environment, we demonstrate that an efficient agent's model explicitly determines its architecture and how much useful work it harvests from the environment. We then show that selecting the maximum-work agent for given environmental data corresponds to finding the maximum-likelihood model. This establishes an equivalence between nonequilibrium thermodynamics and dynamic learning. In this way, work maximization emerges as an organizing principle that underlies learning in adaptive thermodynamic systems.

Oxide heterostructures and superlattices (SLs) have attracted a great deal of attention in recent years owing to the rich exotic properties encountered at their interfaces. We focus on the potential of tunable correlated oxides by investigating the spectral function of the prototypical correlated metal SrVO3, using soft x-ray absorption spectroscopy and resonant inelastic soft x-ray scattering to access both unoccupied and occupied electronic states, respectively. We demonstrate a remarkable level of tunability in the spectral function of SrVO3 by varying its thickness within the SrVO3/SrTiO3 SL, showing that the effects of electron correlation can be tuned from dominating the energy spectrum in a strongly correlated Mott–Hubbard insulator, towards a correlated metal. We show that the effects of dimensionality on the correlated properties of SrVO3 are augmented by interlayer coupling, yielding a highly flexible correlated oxide that may be readily married with other oxide systems.

Objectives

Optical coherence tomography, an imaging modality using near-infrared light, produces cross-sectional tissue images with a lateral pixel resolution of 10 microm. However, normative data is first needed on epithelial thickness for lesion characterisation, and, to date, little exists. The purpose of our study is to measure normal laryngeal epithelial thickness by in vivo optical coherence tomography, and compare these values to those obtained from fixed ex-vivo laryngectomy specimens.

Design and setting

Prospective at a single medical center in California, United States.

Participants

A total of 116 patients undergoing operative endoscopy.

Main outcome measures

Optical coherence tomography images of clinically normal laryngeal subsites were selected. Calibrated measurements of epithelial thickness at various laryngeal subsites were recorded. Measurements of epithelial thickness from corresponding areas were obtained using optical micrometry on histologically normal regions of 15 total laryngectomy specimens. Descriptive statistics were performed.

Results

Mean epithelial optical coherence tomography thicknesses were: true vocal cords (81 microm), false vocal cords (78 microm), subglottis (61 microm), aryepiglottic folds (111 microm), laryngeal epiglottis (116 microm) and lingual epiglottis (170 microm). Epithelial thicknesses in fixed tissues were: true vocal cords (103 microm), false vocal cords (79 microm), aryepiglottic folds (205 microm) subglottis (61 microm), laryngeal epiglottis (38 microm) and lingual epiglottis (130 microm).

Conclusions

Optical coherence tomography does not have the artifacts associated with conventional histologic techniques. The inevitable development of office-based optical coherence tomography devices will increase the precision of laryngeal measurements and contribute to the clinical application of this technology in diagnosing laryngeal disease.

β-Adrenergic signalingpathwaysmediate key aspects of cardiac function. Its dysregulation is associated with a range of cardiac diseases, including dilated cardiomyopathy (DCM). Previously, we established an iPSC model of familial DCM from patients with a mutation in TNNT2, a sarcomeric protein. Here, we found that the β-adrenergic agonist isoproterenol induced mature β-adrenergic signaling in iPSCderived cardiomyocytes (iPSC-CMs) but that this pathway was blunted in DCM iPSC-CMs. Although expression levels of several β-adrenergic signaling components were unaltered between control and DCM iPSC-CMs, we found that phosphodiesterases (PDEs) 2A and PDE3A were upregulated in DCM iPSC-CMs and that PDE2A was also upregulated in DCM patient tissue. We further discovered increased nuclear localization of mutant TNNT2 and epigenetic modifications of PDE genes in both DCM iPSC-CMs and patient tissue. Notably, pharmacologic inhibition of PDE2A and PDE3A restored cAMP levels and ameliorated the impaired β-adrenergic signaling of DCM iPSC-CMs, suggesting therapeutic potential.