Charlotte
We present Charlotte, a framework for composable, authenticated distributed data structures. Charlotte data is stored in blocks that reference each other by hash. Together, all Charlotte blocks form a directed acyclic graph, the blockweb; all observers and applications use subgraphs of the blockweb for their own data structures. Unlike prior systems, Charlotte data structures are composable: applications and data structures can operate fully independently when possible, and share blocks when desired. To support this composability, we define a language-independent format for Charlotte blocks and a network API for Charlotte servers. An authenticated distributed data structure guarantees that data is immutable and self-authenticating: data referenced will be unchanged when it is retrieved. Charlotte extends these guarantees by allowing applications to plug in their own mechanisms for ensuring availability and integrity of data structures. Unlike most traditional distributed systems, including distributed databases, blockchains, and distributed hash tables, Charlotte supports heterogeneous trust: different observers may have their own beliefs about who might fail, and how. Despite heterogeneity of trust, Charlotte presents each observer with a consistent, available view of data. We demonstrate the flexibility of Charlotte by implementing a variety of integrity mechanisms, including consensus and proof of work. We study the power of disentangling availability and integrity mechanisms by building a variety of applications. The results from these examples suggest that developers can use Charlotte to build flexible, fast, composable applications with strong guarantees. …

Multiresolution Graph Attention Network
A large number of deep learning models have been proposed for the text matching problem, which is at the core of various typical natural language processing (NLP) tasks. However, existing deep models are mainly designed for the semantic matching between a pair of short texts, such as paraphrase identification and question answering, and do not perform well on the task of relevance matching between short-long text pairs. This is partially due to the fact that the essential characteristics of short-long text matching have not been well considered in these deep models. More specifically, these methods fail to handle extreme length discrepancy between text pieces and neither can they fully characterize the underlying structural information in long text documents. In this paper, we are especially interested in relevance matching between a piece of short text and a long document, which is critical to problems like query-document matching in information retrieval and web searching. To extract the structural information of documents, an undirected graph is constructed, with each vertex representing a keyword and the weight of an edge indicating the degree of interaction between keywords. Based on the keyword graph, we further propose a Multiresolution Graph Attention Network to learn multi-layered representations of vertices through a Graph Convolutional Network (GCN), and then match the short text snippet with the graphical representation of the document with the attention mechanisms applied over each layer of the GCN. Experimental results on two datasets demonstrate that our graph approach outperforms other state-of-the-art deep matching models. …

mermaid
Generation of diagrams and flowcharts from text in a similar manner as markdown. Ever wanted to simplify documentation and avoid heavy tools like Visio when explaining your code? This is why mermaid was born, a simple markdown-like script language for generating charts from text via javascript. …

Assistive Multi-Armed Bandit
Learning preferences implicit in the choices humans make is a well studied problem in both economics and computer science. However, most work makes the assumption that humans are acting (noisily) optimally with respect to their preferences. Such approaches can fail when people are themselves learning about what they want. In this work, we introduce the assistive multi-armed bandit, where a robot assists a human playing a bandit task to maximize cumulative reward. In this problem, the human does not know the reward function but can learn it through the rewards received from arm pulls; the robot only observes which arms the human pulls but not the reward associated with each pull. We offer sufficient and necessary conditions for successfully assisting the human in this framework. Surprisingly, better human performance in isolation does not necessarily lead to better performance when assisted by the robot: a human policy can do better by effectively communicating its observed rewards to the robot. We conduct proof-of-concept experiments that support these results. We see this work as contributing towards a theory behind algorithms for human-robot interaction. …