During the COVID-19 pandemic, several countries have taken the approach of tiered restrictions which has remained a point of debate due to a lack of transparency. Using the dominance-based rough set approach, we identify patterns in the COVID-19 data pertaining to the UK government's tiered restrictions allocation system. These insights from the analysis are translated into "if-then" type rules, which can easily be interpreted by policy makers. The differences in the rules extracted from different geographical areas suggest inconsistencies in the allocations of tiers in these areas. We found that the differences delineated an overall north south divide in England, however, this divide was driven mostly by London. Based on our analysis, we demonstrate the usefulness of the dominance-based rough sets approach for investigating the fairness and explainabilty of decision making regarding COVID-19 restrictions. The proposed approach and analysis could provide a more transparent approach to localised public health restrictions, which can help ensure greater conformity to the public safety rules.

Chinese Grammatical Error Correction (CGEC) is both a challenging NLP task and a common application in human daily life. Recently, many data-driven approaches are proposed for the development of CGEC research. However, there are two major limitations in the CGEC field: First, the lack of high-quality annotated training corpora prevents the performance of existing CGEC models from being significantly improved. Second, the grammatical errors in widely used test sets are not made by native Chinese speakers, resulting in a significant gap between the CGEC models and the real application. In this paper, we propose a linguistic rules-based approach to construct large-scale CGEC training corpora with automatically generated grammatical errors. Additionally, we present a challenging CGEC benchmark derived entirely from errors made by native Chinese speakers in real-world scenarios. Extensive experiments and detailed analyses not only demonstrate that the training data constructed by our method effectively improves the performance of CGEC models, but also reflect that our benchmark is an excellent resource for further development of the CGEC field.

The Hessian of a differentiable convex function is positive semidefinite. Therefore, checking the Hessian of a given function is a natural approach to certify convexity. However, implementing this approach is not straightforward since it requires a representation of the Hessian that allows its analysis. Here, we implement this approach for a class of functions that is rich enough to support classical machine learning. For this class of functions, it was recently shown how to compute computational graphs of their Hessians. We show how to check these graphs for positive semidefiniteness. We compare our implementation of the Hessian approach with the well-established disciplined convex programming (DCP) approach and prove that the Hessian approach is at least as powerful as the DCP approach for differentiable functions. Furthermore, we show for a state-of-the-art implementation of the DCP approach that, for differentiable functions, the Hessian approach is actually more powerful. That is, it can certify the convexity of a larger class of differentiable functions.

This paper describes our system submitted to Dialogue Robot Competition 2022. Our proposed system is a combined model of rule-based and generation-based dialog systems. The system utilizes HyperCLOVA, a Japanese foundation model, not only to generate responses but also summarization, search information, etc. We also used our original speech recognition system, which was fine-tuned for this dialog task. As a result, our system ranked second in the preliminary round and moved on to the finals.

The high-dimesionality, non-linearity and emergent properties of complex systems pose a challenge to identifying general laws in the same manner that has been so successful in simpler physical systems. In Anderson's seminal work on why "more is different" he pointed to how emergent, macroscale patterns break symmetries of the underlying microscale laws. Yet, less recognized is that these large-scale, emergent patterns must also retain some symmetries of the microscale rules. Here we introduce a new, relational macrostate theory (RMT) that defines macrostates in terms of symmetries between two mutually predictive observations, and develop a machine learning architecture, MacroNet, that identifies macrostates. Using this framework, we show how macrostates can be identifed across systems ranging in complexity from the simplicity of the simple harmonic oscillator to the much more complex spatial patterning characteristic of Turing instabilities. Furthermore, we show how our framework can be used for the inverse design of microstates consistent with a given macroscopic property -- in Turing patterns this allows us to design underlying rule with a given specification of spatial patterning, and to identify which rule parameters most control these patterns. By demonstrating a general theory for how macroscopic properties emerge from conservation of symmetries in the mapping between observations, we provide a machine learning framework that allows a unified approach to identifying macrostates in systems from the simple to complex, and allows the design of new examples consistent with a given macroscopic property.

Machine learning (ML) formalizes the problem of getting computers to learn from experience as optimization of performance according to some metric(s) on a set of data examples. This is in contrast to requiring behaviour specified in advance (e.g. by hard-coded rules). Formalization of this problem has enabled great progress in many applications with large real-world impact, including translation, speech recognition, self-driving cars, and drug discovery. But practical instantiations of this formalism make many assumptions - for example, that data are i.i.d.: independent and identically distributed - whose soundness is seldom investigated. And in making great progress in such a short time, the field has developed many norms and ad-hoc standards, focused on a relatively small range of problem settings. As applications of ML, particularly in artificial intelligence (AI) systems, become more pervasive in the real world, we need to critically examine these assumptions, norms, and problem settings, as well as the methods that have become de-facto standards. There is much we still do not understand about how and why deep networks trained with stochastic gradient descent are able to generalize as well as they do, why they fail when they do, and how they will perform on out-of-distribution data. In this thesis I cover some of my work towards better understanding deep net generalization, identify several ways assumptions and problem settings fail to generalize to the real world, and propose ways to address those failures in practice.

Ten years ago, 12-year-old Rory Staunton dove for a ball in gym class and scraped his arm. He woke up the next day with a 104 F fever, so his parents took him to the pediatrician and eventually the emergency room. It was just the stomach flu, they were told. Three days later, Rory died of sepsis after bacteria from the scrape infiltrated his blood and triggered organ failure. "How does that happen in a modern society?" his father, Ciaran Staunton, said in a recent interview with Undark. Each year in the United States, sepsis kills over a quarter million people -- more than stroke, diabetes, or lung cancer.

Natural language processing (NLP) is a powerful combination of linguistics and computer science that, through the study of language and the creation of intelligent systems, makes human language as intelligible to machines as it would be for a human being, whether in text or speech format. As a branch of artificial intelligence (AI), NLP enables computers and machines to understand, interpret and manipulate human language using computational linguistics and statistical models, machine learning methods and deep learning processes. The knowledge extracted by these technologies is converted into algorithms that teach machines to perform a myriad of tasks that are infinitely valuable to businesses. The more data NLP algorithms receive, the more precise text analysis models become. NLP includes an immense diversity of techniques, from statistical and machine learning methods to algorithmic and rule-based approaches.

We present POTATO, a task- and languageindependent framework for human-in-the-loop (HITL) learning of rule-based text classifiers using graph-based features. POTATO handles any type of directed graph and supports parsing text into Abstract Meaning Representations (AMR), Universal Dependencies (UD), and 4lang semantic graphs. A streamlit-based user interface allows users to build rule systems from graph patterns, provides real-time evaluation based on ground truth data, and suggests rules by ranking graph features using interpretable machine learning models. Users can also provide patterns over graphs using regular expressions, and POTATO can recommend refinements of such rules. POTATO is applied in projects across domains and languages, including classification tasks on German legal text and English social media data. All components of our system are written in Python, can be installed via pip, and are released under an MIT License on GitHub.

Call Graph evolution analytics can aid a software engineer when maintaining or evolving a software system. This paper proposes Call Graph Evolution Analytics to extract information from an evolving call graph ECG = CG_1, CG_2,... CG_N for their version series VS = V_1, V_2, ... V_N of an evolving software system. This is done using Call Graph Evolution Rules (CGERs) and Call Graph Evolution Subgraphs (CGESs). Similar to association rule mining, the CGERs are used to capture co-occurrences of dependencies in the system. Like subgraph patterns in a call graph, the CGESs are used to capture evolution of dependency patterns in evolving call graphs. Call graph analytics on the evolution in these patterns can identify potentially affected dependencies (or procedure calls) that need attention. The experiments are done on the evolving call graphs of 10 large evolving systems to support dependency evolution management. We also consider results from a detailed study for evolving call graphs of Maven-Core's version series.