Big Data Analytics and Deep Learning are two high-focus of data science. Big Data has become important as many organizations both public and private have been collecting massive amounts of domain-specific information, which can contain useful information about problems such as national intelligence, cyber security, fraud detection, marketing, and medical informatics. Companies such as Google and Microsoft are analyzing large volumes of data for business analysis and decisions, impacting existing and future technology. Deep Learning algorithms extract high-level, complex abstractions as data representations through a hierarchical learning process. Complex abstractions are learnt at a given level based on relatively simpler abstractions formulated in the preceding level in the hierarchy. A key benefit of Deep Learning is the analysis and learning of massive amounts of unsupervised data, making it a valuable tool for Big Data Analytics where raw data is largely unlabeled and un-categorized. In the present study, we explore how Deep Learning can be utilized for addressing some important problems in Big Data Analytics, including extracting complex patterns from massive volumes of data, semantic indexing, data tagging, fast information retrieval, and simplifying discriminative tasks. We also investigate some aspects of Deep Learning research that need further exploration to incorporate specific challenges introduced by Big Data Analytics, including streaming data, high-dimensional data, scalability of models, and distributed computing. We conclude by presenting insights into relevant future works by posing some questions, including defining data sampling criteria, domain adaptation modeling, defining criteria for obtaining useful data abstractions, improving semantic indexing, semi-supervised learning, and active learning.

/pdf/deep-learning-applications-and-challenges-in-big-data-oblfhz8zye.pdf

Deep learning applications and challenges in big data analytics

The Creative Mind: Myths and Mechanisms.

We introduce a novel neural network architecture for encoding and synthesis of 3D shapes, particularly their structures. Our key insight is that 3D shapes are effectively characterized by their hierarchical organization of parts, which reflects fundamental intra-shape relationships such as adjacency and symmetry. We develop a recursive neural net (RvNN) based autoencoder to map a flat, unlabeled, arbitrary part layout to a compact code. The code effectively captures hierarchical structures of man-made 3D objects of varying structural complexities despite being fixed-dimensional: an associated decoder maps a code back to a full hierarchy. The learned bidirectional mapping is further tuned using an adversarial setup to yield a generative model of plausible structures, from which novel structures can be sampled. Finally, our structure synthesis framework is augmented by a second trained module that produces fine-grained part geometry, conditioned on global and local structural context, leading to a full generative pipeline for 3D shapes. We demonstrate that without supervision, our network learns meaningful structural hierarchies adhering to perceptual grouping principles, produces compact codes which enable applications such as shape classification and partial matching, and supports shape synthesis and interpolation with significant variations in topology and geometry.

https://dl.acm.org/doi/pdf/10.1145/3072959.3073637

GRASS: generative recursive autoencoders for shape structures

Iteration can help people improve ideas It can also give rise to fixation, continuously refining one option without considering others Does creating and receiving feedback on multiple prototypes in parallel, as opposed to serially, affect learning, self-efficacy, and design explorationq An experiment manipulated whether independent novice designers created graphic Web advertisements in parallel or in series Serial participants received descriptive critique directly after each prototype Parallel participants created multiple prototypes before receiving feedback As measured by click-through data and expert ratings, ads created in the Parallel condition significantly outperformed those from the Serial condition Moreover, independent raters found Parallel prototypes to be more diverse Parallel participants also reported a larger increase in task-specific self-confidence This article outlines a theoretical foundation for why parallel prototyping produces better design results and discusses the implications for design education

/pdf/parallel-prototyping-leads-to-better-design-results-more-3e10rkuoif.pdf

Parallel prototyping leads to better design results, more divergence, and increased self-efficacy

Data-driven models help mobile app designers understand best practices and trends, and can be used to make predictions about design performance and support the creation of adaptive UIs. This paper presents Rico, the largest repository of mobile app designs to date, created to support five classes of data-driven applications: design search, UI layout generation, UI code generation, user interaction modeling, and user perception prediction. To create Rico, we built a system that combines crowdsourcing and automation to scalably mine design and interaction data from Android apps at runtime. The Rico dataset contains design data from more than 9.7k Android apps spanning 27 categories. It exposes visual, textual, structural, and interactive design properties of more than 72k unique UI screens. To demonstrate the kinds of applications that Rico enables, we present results from training an autoencoder for UI layout similarity, which supports query- by-example search over UIs.

https://dl.acm.org/doi/pdf/10.1145/3126594.3126651

Rico: A Mobile App Dataset for Building Data-Driven Design Applications

Outfits in online fashion data are composed of items of many different types (e.g. top, bottom, shoes) that share some stylistic relationship with one another. A representation for building outfits requires a method that can learn both notions of similarity (for example, when two tops are interchangeable) and compatibility (items of possibly different type that can go together in an outfit). This paper presents an approach to learning an image embedding that respects item type, and jointly learns notions of item similarity and compatibility in an end-to-end model. To evaluate the learned representation, we crawled 68,306 outfits created by users on the Polyvore website. Our approach obtains 3–5% improvement over the state-of-the-art on outfit compatibility prediction and fill-in-the-blank tasks using our dataset, as well as an established smaller dataset, while supporting a variety of useful queries (Code and data: https://github.com/mvasil/fashion-compatibility).

/pdf/learning-type-aware-embeddings-for-fashion-compatibility-klnprn67rb.pdf

Learning Type-Aware Embeddings for Fashion Compatibility

Designers often use examples for inspiration; examples offer contextualized instances of how form and content integrate. Can interactive example galleries bring this practice to everyday users doing design work, and does working with examples help the designs they create? This paper explores whether people can realize significant value from explicit mechanisms for designing by example modification. We present the results of three studies, finding that independent raters prefer designs created with the aid of examples, that examples may benefit novices more than experienced designers, that users prefer adaptively selected examples to random ones, and that users make use of multiple examples when creating new designs. To enable these studies and demonstrate how software tools can facilitate designing with examples, we introduce interface techniques for browsing and borrowing from a corpus of examples, manifest in the Adaptive Ideas Web design tool. Adaptive Ideas leverages a faceted metadata interface for viewing and navigating example galleries.

/pdf/designing-with-interactive-example-galleries-2uqcx4u0i4.pdf

Designing with interactive example galleries

The Web provides a corpus of design examples unparalleled in human history. However, leveraging existing designs to produce new pages is often difficult. This paper introduces the Bricolage algorithm for transferring design and content between Web pages. Bricolage employs a novel, structured-prediction technique that learns to create coherent mappings between pages by training on human-generated exemplars. The produced mappings are then used to automatically transfer the content from one page into the style and layout of another. We show that Bricolage can learn to accurately reproduce human page mappings, and that it provides a general, efficient, and automatic technique for retargeting content between a variety of real Web pages.

Bricolage: example-based retargeting for web design

Advances in data mining and knowledge discovery have transformed the way Web sites are designed. However, while visual presentation is an intrinsic part of the Web, traditional data mining techniques ignore render-time page structures and their attributes. This paper introduces design mining for the Web: using knowledge discovery techniques to understand design demographics, automate design curation, and support data-driven design tools. This idea is manifest in Webzeitgeist, a platform for large-scale design mining comprising a repository of over 100,000 Web pages and 100 million design elements. This paper describes the principles driving design mining, the implementation of the Webzeitgeist architecture, and the new class of data-driven design applications it enables.

Ranjitha Kumar

Papers

Rico: A Mobile App Dataset for Building Data-Driven Design Applications

Learning Type-Aware Embeddings for Fashion Compatibility

Designing with interactive example galleries

Bricolage: example-based retargeting for web design

Webzeitgeist: design mining the web