Proof by induction is a long-standing challenge in Computer Science. Induction tactics of proof assistants facilitate proof by induction, but rely on humans to manually specify how to apply induction. In this paper, we present SeLFiE, a domain-specific language to encode experienced users' expertise on how to apply the induct tactic in Isabelle/HOL: when we apply an induction heuristic written in SeLFiE to an inductive problem and arguments to the induct tactic, the SeLFiE interpreter examines both the syntactic structure of the problem and semantics of the relevant constants to judge whether the arguments to the induct tactic are plausible according to the heuristic. Then, we present semantic_induct, an automatic tool to recommend how to apply the induct tactic. Given an inductive problem, semantic_induct produces candidate arguments to the induct tactic and selects promising ones using heuristics written in SeLFiE. Our evaluation based on 254 inductive problems from nine problem domains show that semantic_induct achieved 15.7 percentage points of improvements in coincidence rates for the three most promising recommendations while achieving 43% of reduction in the median value for the execution time when compared to an existing tool, smart_induct.

When an apparently indecipherable manuscript from a lost language turns up, AI can help. But first, how is a language born and how does it die (or get lost)? We really don't know how human language was born; theories abound but all we know for sure is that it is unique. In a 2017 paper at BMC Biology, evolutionary biologist Mark Pagel states flatly, "Human language is unique among all forms of animal communication." Most ape sign language, for example, is concerned with requests for food.

When the elephant arrived in the night, on the hunt for sugarcane, Uthorn Kanthong was waiting for him. Like many of his neighbors, the 69-year-old Thai farmer had taken to staying in his fields into the late hours, to try and scare off elephants that came to snack on his crop. He usually returned home by midnight. But that night in 2018, he didn't come back. Worried, his daughter sent out family and friends to look for him.

Technology is invading in every sector. New inventions, innovation and devices are making life easier for everyone. Voice recognition technology is one such amazing initiative to look for in the growing innovation era. Voice recognition also known as speech recognition, is a computer software program or a hardware device with the ability to receive, interpreting and understanding voice and carry out commands. The technology unravels the feature to easily create and control documents by speaking, with the help of technology.

Of a plethora of digital technologies running around in the web space and an immense part of the transportation industry; AI or artificial intelligence has something significant to offer. It is one such technology which is building concrete impact all over the market place, especially with the truck rental services. Perhaps, there are millions of freight rental providers all over the globe, and more than half are already accustomed to the new end technology. Even the collision of AI and supply chain has given proven results. To know the topic better, it would be a good idea, to begin with, a conversation right now.

Natural language processing systems often struggle with out-of-vocabulary (OOV) terms, which do not appear in training data. Blends, such as "innoventor", are one particularly challenging class of OOV, as they are formed by fusing together two or more bases that relate to the intended meaning in unpredictable manners and degrees. In this work, we run experiments on a novel dataset of English OOV blends to quantify the difficulty of interpreting the meanings of blends by large-scale contextual language models such as BERT. We first show that BERT's processing of these blends does not fully access the component meanings, leaving their contextual representations semantically impoverished. We find this is mostly due to the loss of characters resulting from blend formation. Then, we assess how easily different models can recognize the structure and recover the origin of blends, and find that context-aware embedding systems outperform character-level and context-free embeddings, although their results are still far from satisfactory.

Graph-based Semi-Supervised Learning (SSL) aims to transfer the labels of a handful of labeled data to the remaining massive unlabeled data via a graph. As one of the most popular graph-based SSL approaches, the recently proposed Graph Convolutional Networks (GCNs) have gained remarkable progress by combining the sound expressiveness of neural networks with graph structure. Nevertheless, the existing graph-based methods do not directly address the core problem of SSL, i.e., the shortage of supervision, and thus their performances are still very limited. To accommodate this issue, a novel GCN-based SSL algorithm is presented in this paper to enrich the supervision signals by utilizing both data similarities and graph structure. Firstly, by designing a semi-supervised contrastive loss, improved node representations can be generated via maximizing the agreement between different views of the same data or the data from the same class. Therefore, the rich unlabeled data and the scarce yet valuable labeled data can jointly provide abundant supervision information for learning discriminative node representations, which helps improve the subsequent classification result. Secondly, the underlying determinative relationship between the data features and input graph topology is extracted as supplementary supervision signals for SSL via using a graph generative loss related to the input features. Intensive experimental results on a variety of real-world datasets firmly verify the effectiveness of our algorithm compared with other state-of-the-art methods.

Exam cheating is as ancient a practice as education itself. Back in ancient China, cheating on the Imperial exams was a serious offense. And yet a Qing dynasty cheatsheet--in the shape of a handkerchief with 10,000 symbols in microscopic writing--is on display at the Minneapolis Institute of Arts, showing that students have always been inclined to borrow knowledge and ideas. Exam cheating is as ancient a practice as education itself. Historically, students could attempt to use their own handwriting as false proof of originality, but now the digital age has changed that.

Structured belief states are crucial for user goal tracking and database query in task-oriented dialog systems. However, training belief trackers often requires expensive turn-level annotations of every user utterance. In this paper we aim at alleviating the reliance on belief state labels in building end-to-end dialog systems, by leveraging unlabeled dialog data towards semi-supervised learning. We propose a probabilistic dialog model, called the LAtent BElief State (LABES) model, where belief states are represented as discrete latent variables and jointly modeled with system responses given user inputs. Such latent variable modeling enables us to develop semi-supervised learning under the principled variational learning framework. Furthermore, we introduce LABES-S2S, which is a copy-augmented Seq2Seq model instantiation of LABES. In supervised experiments, LABES-S2S obtains strong results on three benchmark datasets of different scales. In utilizing unlabeled dialog data, semi-supervised LABES-S2S significantly outperforms both supervised-only and semi-supervised baselines. Remarkably, we can reduce the annotation demands to 50% without performance loss on MultiWOZ.

Given a stream of entries in a multi-aspect data setting i.e., entries having multiple dimensions, how can we detect anomalous activities? For example, in the intrusion detection setting, existing work seeks to detect anomalous events or edges in dynamic graph streams, but this does not allow us to take into account additional attributes of each entry. Our work aims to define a streaming multi-aspect data anomaly detection framework, termed MStream, which can detect unusual group anomalies as they occur, in a dynamic manner. MStream has the following properties: (a) it detects anomalies in multi-aspect data including both categorical and numeric attributes; (b) it is online, thus processing each record in constant time and constant memory; (c) it can capture the correlation between multiple aspects of the data. MStream is evaluated over the KDDCUP99, CICIDS-DoS, UNSW-NB 15 and CICIDS-DDoS datasets, and outperforms state-of-the-art baselines.