Product search is the most common way for people to satisfy their shopping needs on e-commerce websites. Products are typically annotated with one of several broad categorical tags, such as "Clothing" or "Electronics", as well as finer-grained categories like "Refrigerator" or "TV", both under "Electronics". These tags are used to construct a hierarchy of query categories. Feature distributions such as price and brand popularity vary wildly across query categories. In addition, feature importance for the purpose of CTR/CVR predictions differs from one category to another. In this work, we leverage the Mixture of Expert (MoE) framework to learn a ranking model that specializes for each query category. In particular, our gate network relies solely on the category ids extracted from the user query. While classical MoE's pick expert towers spontaneously for each input example, we explore two techniques to establish more explicit and transparent connections between the experts and query categories. To help differentiate experts on their domain specialties, we introduce a form of adversarial regularization among the expert outputs, forcing them to disagree with one another. As a result, they tend to approach each prediction problem from different angles, rather than copying one another. This is validated by a much stronger clustering effect of the gate output vectors under different categories. In addition, soft gating constraints based on the categorical hierarchy are imposed to help similar products choose similar gate values. and make them more likely to share similar experts. This allows aggregation of training data among smaller sibling categories to overcome data scarcity issues among the latter. Experiments on a learning-to-rank dataset gathered from a leading e-commerce search log demonstrate that MoE with our improvements consistently outperforms competing models.

Usually, you pop up in an exhibition, coming from vivid streets to the Silent Hall of art. The exhibition pops up where you are, suddenly, amidst your next Zoom call, or while you are checking your emails. And you are exposed to it. In these crazy pandemic times, they found a perfect way to present art, without put the visitors in danger to be Corona'ed: Plug-In. You install a plug-in to your browser, and every hour another artwork overfloods your PC windows.

In an attempt to automate industrial designing, researchers from Princeton University and Columbia University introduced a large dataset of 15 million two-dimensional real-world computer-aided designs -- SketchGraphs. Along with that to facilitate research in ML-aided design, they also launched an open-source data processing pipeline. Introduced during the International Conference on Machine Learning, SketchGraphs is aimed to train the artificial intelligence machine with this large dataset, in order to expertise it to assist humans in creating CAD models. In a recent paper, researchers revealed that each of the CAD sketches is represented with a geometric constraint graph and the understanding of the line and shape sequence in which the design was initially created. This will enable the predictions of what is going to be designed next.

Parametric computer-aided design (CAD) is the dominant paradigm in mechanical engineering for physical design. Distinguished by relational geometry, parametric CAD models begin as two-dimensional sketches consisting of geometric primitives (e.g., line segments, arcs) and explicit constraints between them (e.g., coincidence, perpendicularity) that form the basis for three-dimensional construction operations. Training machine learning models to reason about and synthesize parametric CAD designs has the potential to reduce design time and enable new design workflows. Additionally, parametric CAD designs can be viewed as instances of constraint programming and they offer a well-scoped test bed for exploring ideas in program synthesis and induction. To facilitate this research, we introduce SketchGraphs, a collection of 15 million sketches extracted from real-world CAD models coupled with an open-source data processing pipeline.

The study of phase transition behaviour in SAT has led to deeper understanding and algorithmic improvements of modern SAT solvers. Motivated by these prior studies of phase transitions in SAT, we seek to study the behaviour of size and compile-time behaviour for random k-CNF formulas in the context of knowledge compilation. We perform a rigorous empirical study and analysis of the size and runtime behavior for different knowledge compilation forms (and their corresponding compilation algorithms): d-DNNFs, SDDs and OBDDs across multiple tools and compilation algorithms. We employ instances generated from the random k-CNF model with varying generation parameters to empirically reason about the expected and median behavior of size and compilation-time for these languages. Our work is similar in spirit to the early work in CSP community on phase transition behavior in SAT/CSP. In a similar spirit, we identify the interesting behavior with respect to different parameters: clause density and solution density, a novel control parameter that we identify for the study of phase transition behavior in the context of knowledge compilation. Furthermore, we summarize our empirical study in terms of two concrete conjectures; a rigorous study of these conjectures will possibly require new theoretical tools.

Modern software deployment process produces software that is uniform, and hence vulnerable to large-scale code-reuse attacks. Compiler-based diversification improves the resilience and security of software systems by automatically generating different assembly code versions of a given program. Existing techniques are efficient but do not have a precise control over the quality of the generated code variants. This paper introduces Diversity by Construction (DivCon), a constraint-based compiler approach to software diversification. Unlike previous approaches, DivCon allows users to control and adjust the conflicting goals of diversity and code quality. A key enabler is the use of Large Neighborhood Search (LNS) to generate highly diverse assembly code efficiently. Experiments using two popular compiler benchmark suites confirm that there is a trade-off between quality of each assembly code version and diversity of the entire pool of versions. Our results show that DivCon allows users to trade between these two properties by generating diverse assembly code for a range of quality bounds. In particular, the experiments show that DivCon is able to mitigate code-reuse attacks effectively while delivering near-optimal code (< 10% optimality gap). For constraint programming researchers and practitioners, this paper demonstrates that LNS is a valuable technique for finding diverse solutions. For security researchers and software engineers, DivCon extends the scope of compiler-based diversification to performance-critical and resource-constrained applications.

What do self-driving cars, face recognition, web search, industrial robots, missile guidance, and tumor detection have in common? They are all complex real world problems being solved with applications of intelligence (AI). This course will provide a broad understanding of the basic techniques for building intelligent computer systems and an understanding of how AI is applied to problems. You will learn about the history of AI, intelligent agents, state-space problem representations, uninformed and heuristic search, game playing, logical agents, and constraint satisfaction problems. Hands on experience will be gained by building a basic search agent.

Parametric computer-aided design (CAD) is the dominant paradigm in mechanical engineering for physical design. Distinguished by relational geometry, parametric CAD models begin as two-dimensional sketches consisting of geometric primitives (e.g., line segments, arcs) and explicit constraints between them (e.g., coincidence, perpendicularity) that form the basis for three-dimensional construction operations. Training machine learning models to reason about and synthesize parametric CAD designs has the potential to reduce design time and enable new design workflows. Additionally, parametric CAD designs can be viewed as instances of constraint programming and they offer a well-scoped test bed for exploring ideas in program synthesis and induction. To facilitate this research, we introduce SketchGraphs, a collection of 15 million sketches extracted from real-world CAD models coupled with an open-source data processing pipeline. Each sketch is represented as a geometric constraint graph where edges denote designer-imposed geometric relationships between primitives, the nodes of the graph. We demonstrate and establish benchmarks for two use cases of the dataset: generative modeling of sketches and conditional generation of likely constraints given unconstrained geometry.

During the virtually held Robotics: Science and Systems 2020 conference this week, scientists affiliated with the National University of Singapore (NUS) presented research that combines robotic vision and touch sensing with Intel-designed neuromorphic processors. The researchers claim the "electronic skin" -- dubbed Asynchronous Coded Electronic Skin (ACES) -- can detect touches more than 1,000 times faster than the human nervous system and identify the shape, texture, and hardness of objects within 10 milliseconds. At the same time, ACES is designed to be modular and highly robust to damage, ensuring it can continue functioning as long as at least one sensor remains. The human sense of touch is fine-grained enough to distinguish between surfaces that differ by only a single layer of molecules, yet the majority of today's autonomous robots operate solely via visual, spatial, and inertial processing techniques. Bringing humanlike touch to machines could significantly improve their utility and even lead to new use cases.

In this paper, we focus on qualitative temporal sequences of topological information. We firstly consider the context of topological temporal sequences of length greater than 3 describing the evolution of regions at consecutive time points. We show that there is no Cartesian subclass containing all the basic relations and the universal relation for which the algebraic closure decides satisfiability. However, we identify some tractable subclasses, by giving up the relations containing the non-tangential proper part relation and not containing the tangential proper part relation. We then formalize an alternative semantics for temporal sequences. We place ourselves in the context of the topological temporal sequences describing the evolution of regions on a partition of time (i.e. an alternation of instants and intervals). In this context, we identify large tractable fragments.