Organic small molecule semiconductors have many advantages over their polymer analogues. However, to fabricate organic semiconductor-based devices using solution processing, it is requisite to eliminate dewetting to ensure film uniformity and desirable to assemble nanoscopic features with tailored macroscopic alignment without compromising their electronic properties. To this end, we present a modular supramolecular approach. A quaterthiophene organic semiconductor is attached to the side chains of poly(4-vinylpyridine) via noncovalent hydrogen bonds to form supramolecular assemblies that act as p-type semiconductors in field-effect transistors. In thin films, the quaterthiophenes can be readily assembled into microdomains, tens of nanometers in size, oriented normal to the surface. The supramolecules exhibited the same field-effect mobilities as that of the quaterthiophene alone (10(-4) cm(2)/(V.s)). Since the organic semiconductors can be readily substituted, this modular supramolecular approach is a viable method for the fabrication of functional, nanostructured organic semiconductor films using solution processing.
We describe an innovative and scalable recommendation system successfully deployed at eBay. To build recommenders for long-tail marketplaces requires projection of volatile items into a persistent space of latent products. We first present a generative clustering model for collections of unstructured, heterogeneous, and ephemeral item data, under the assumption that items are generated from latent products. An item is represented as a vector of independently and distinctly distributed variables, while a latent product is characterized as a vector of probability distributions, respectively. The probability distributions are chosen as natural stochastic models for different types of data. The learning objective is to maximize the total intra-cluster coherence measured by the sum of log likelihoods of items under such a generative process. In the space of latent products, robust recommendations can then be derived using naive Bayes for ranking, from historical transactional data. Item-based recommendations are achieved by inferring latent products from unseen items. In particular, we develop a probabilistic scoring function of recommended items, which takes into account item-product membership, product purchase probability, and the important auction-end-time factor. With the holistic probabilistic measure of a prospective item purchase, one can further maximize the expected revenue and the more subjective user satisfaction as well. We evaluated the latent product clustering and recommendation ranking models using real-world e-commerce data from eBay, in both forms of offline simulation and online A/B testing. In the recent production launch, our system yielded 3-5 folds improvement over the existing production system in click-through, purchase-through and gross merchandising value; thus now driving 100% related recommendation traffic with billions of items at eBay. We believe that this work provides a practical yet principled framework for recommendation in the domains with affluent user self-input data.
Convolutional image classifiers can achieve high predictive accuracy, but quantifying their uncertainty remains an unresolved challenge, hindering their deployment in consequential settings. Existing uncertainty quantification techniques, such as Platt scaling, attempt to calibrate the network's probability estimates, but they do not have formal guarantees. We present an algorithm that modifies any classifier to output a predictive set containing the true label with a user-specified probability, such as 90%. The algorithm is simple and fast like Platt scaling, but provides a formal finite-sample coverage guarantee for every model and dataset. Our method modifies an existing conformal prediction algorithm to give more stable predictive sets by regularizing the small scores of unlikely classes after Platt scaling. In experiments on both Imagenet and Imagenet-V2 with ResNet-152 and other classifiers, our scheme outperforms existing approaches, achieving coverage with sets that are often factors of 5 to 10 smaller than a stand-alone Platt scaling baseline.
Given the need to provide users with reasonable feedback about the "costs" their network usage incurs, and the increasingly commercial nature of the Internet, we believe that the allocation of cost among users will play an important role in future networks. This paper discusses cost allocation in the context of multicast flows. The question we discuss is this: when a single data flow is shared among many receivers, how does one split the cost of that flow among the receivers? Multicast routing increases network efficiency by using a single shared delivery tree. We address the issue of how these savings are allocated among the various members of the multicast group. We first consider an axiomatic approach to the problem, analyzing the implications of different distributive notions on the resulting allocations. We then consider a one-pass mechanism to implement such allocation schemes and investigate the family of allocation schemes such mechanisms can support.
