Abstract-An increasing number of real-world applications are associated with streaming data drawn from drifting and nonstationary distributions that change over time. These applications demand new algorithms that can learn and adapt to such changes, also known as concept drift. Proper characterization of such data with existing approaches typically requires substantial amount of labeled instances, which may be difficult, expensive, or even impractical to obtain. In this paper, we introduce compacted object sample extraction (COMPOSE), a computational geometry-based framework to learn from nonstationary streaming data, where labels are unavailable (or presented very sporadically) after initialization. We introduce the algorithm in detail, and discuss its results and performances on several synthetic and real-world data sets, which demonstrate the ability of the algorithm to learn under several different scenarios of initially labeled streaming environments. On carefully designed synthetic data sets, we compare the performance of COMPOSE against the optimal Bayes classifier, as well as the arbitrary subpopulation tracker algorithm, which addresses a similar environment referred to as extreme verification latency. Furthermore, using the real-world National Oceanic and Atmospheric Administration weather data set, we demonstrate that COMPOSE is competitive even with a well-established and fully supervised nonstationary learning algorithm that receives labeled data in every batch.Index Terms-Alpha shape, concept drift, nonstationary environment, semisupervised learning (SSL), verification latency.

Low back pain is a major public health issue associated with degeneration of the intervertebral disc (IVD). The early stages of degeneration are characterized by the dehydration of the central, gelatinous portion of the IVD, the nucleus pulposus (NP). One possible treatment approach is to replace the NP in the early stages of IVD degeneration with a hydrogel that restores healthy biomechanics while supporting tissue regeneration. The present study evaluates a novel thermosensitive hydrogel based on poly(N‐isopropylacrylamide‐graft‐chondroitin sulfate) (PNIPAAM‐g‐CS) for NP replacement. The hypothesis was tested that the addition of freeze‐dried, calcium crosslinked alginate microparticles (MPs) to aqueous solutions of PNIPAAm‐g‐CS would enable tuning of the rheological properties of the injectable solution, as well as the bioadhesive and mechanical properties of the thermally precipitated composite gel. Further, we hypothesized that the composite would support encapsulated cell viability and differentiation. Structure‐material property relationships were evaluated by varying MP concentration and diameter. The addition of high concentrations (50 mg/mL) of small MPs (20 ± 6 μm) resulted in the greatest improvement in injectability, compressive mechanical properties, and bioadhesive strength of PNIPAAm‐g‐CS. This combination of PNIPAAM‐g‐CS and alginate MPs supported the survival, proliferation, and differentiation of adipose derived mesenchymal stem cells toward an NP‐like phenotype in the presence of soluble GDF‐6. When implanted ex vivo into the intradiscal cavity of degenerated porcine IVDs, the formulation restored the compressive and neutral zone stiffnesses to intact values and resisted expulsion under lateral bending. Overall, results indicate the potential of the hydrogel composite to serve as a scaffold for supporting NP regeneration. This work uniquely demonstrates that encapsulation of re‐hydrating polysaccharide‐based MPs may be an effective method for improving key functional properties of in situ forming hydrogels for orthopedic tissue engineering applications.

Abstract-Semi-supervised learning (SSL) in non-stationary environments has received relatively little attention in machine learning, despite a growing number of applications that can benefit from a properly configured SSL algorithm. Previous works in learning non-stationary data have analyzed such cases where both labeled and unlabeled instances are received at every time step and/or in regular intervals; however, to the best of our knowledge, no work has investigated the case where labeled instances are received only at the initial time step, followed by unlabeled instances provided in subsequent time steps. In this proof-of-concept work, we propose a new framework for learning in a non-stationary environment that provides only unlabeled data after the initial time step, to which we refer to as initially labeled environment. The proposed framework generates labels for previously unlabeled data at each time step to be combined with incoming unlabeled data -possibly from a drifting distribution -using a compacted polytope sample extraction algorithm. We have conducted two experiments to demonstrate the feasibility and reliability of the approach. This proof-of-concept is presented in two dimensions; however, the algorithm can be extended to higher dimensions with appropriate modifications.

No abstract