arXiv subjects

Kim Nguyen

Publications and source records attributed to Kim Nguyen.

7 recordsLinked to original sources

arXiv2025

SHREC 2025: Retrieval of Optimal Objects for Multi-modal Enhanced Language and Spatial Assistance (ROOMELSA)

Recent 3D retrieval systems are typically designed for simple, controlled scenarios, such as identifying an object from a cropped image or a brief description. However, real-world scenarios are more complex, often requiring the recognition of an object in a cluttered scene based on a vague, free-form description. To this end, we present ROOMELSA, a new benchmark designed to evaluate a system's ability to interpret natural language. Specifically, ROOMELSA attends to a specific region within a panoramic room image and accurately retrieves the corresponding 3D model from a large database. In addition, ROOMELSA includes over 1,600 apartment scenes, nearly 5,200 rooms, and more than 44,000 targeted queries. Empirically, while coarse object retrieval is largely solved, only one top-performing model consistently ranked the correct match first across nearly all test cases. Notably, a lightweight CLIP-based model also performed well, although it struggled with subtle variations in materials, part structures, and contextual cues, resulting in occasional errors. These findings highlight the importance of tightly integrating visual and language understanding. By bridging the gap between scene-level grounding and fine-grained 3D retrieval, ROOMELSA establishes a new benchmark for advancing robust, real-world 3D recognition systems.

cs.CV↗

arXiv2023

Polymorphic Type Inference for Dynamic Languages

We present a type system that combines, in a controlled way, first-order polymorphism with intersectiontypes, union types, and subtyping, and prove its safety. We then define a type reconstruction algorithm that issound and terminating. This yields a system in which unannotated functions are given polymorphic types(thanks to Hindley-Milner) that can express the overloaded behavior of the functions they type (thanks tothe intersection introduction rule) and that are deduced by applying advanced techniques of type narrowing(thanks to the union elimination rule). This makes the system a prime candidate to type dynamic languages.

cs.PL↗

arXiv2019

Revisiting Occurrence Typing

We revisit occurrence typing, a technique to refine the type of variables occurring in type-cases and, thus, capturesome programming patterns used in untyped languages. Although occurrence typing was tied from its inceptionto set-theoretic types-union types, in particular-it never fully exploited the capabilities of these types. Here weshow how, by using set-theoretic types, it is possible to develop a general typing framework that encompasses andgeneralizes several aspects of current occurrence typing proposals and that can be applied to tackle other problemssuch as the reconstruction of intersection types for unannotated or partially annotated functions and the optimizationof the compilation of gradually typed languages.

cs.PL↗

arXiv2017

DeepRadiologyNet: Radiologist Level Pathology Detection in CT Head Images

We describe a system to automatically filter clinically significant findings from computerized tomography (CT) head scans, operating at performance levels exceeding that of practicing radiologists. Our system, named DeepRadiologyNet, builds on top of deep convolutional neural networks (CNNs) trained using approximately 3.5 million CT head images gathered from over 24,000 studies taken from January 1, 2015 to August 31, 2015 and January 1, 2016 to April 30 2016 in over 80 clinical sites. For our initial system, we identified 30 phenomenological traits to be recognized in the CT scans. To test the system, we designed a clinical trial using over 4.8 million CT head images (29,925 studies), completely disjoint from the training and validation set, interpreted by 35 US Board Certified radiologists with specialized CT head experience. We measured clinically significant error rates to ascertain whether the performance of DeepRadiologyNet was comparable to or better than that of US Board Certified radiologists. DeepRadiologyNet achieved a clinically significant miss rate of 0.0367% on automatically selected high-confidence studies. Thus, DeepRadiologyNet enables significant reduction in the workload of human radiologists by automatically filtering studies and reporting on the high-confidence ones at an operating point well below the literal error rate for US Board Certified radiologists, estimated at 0.82%.

cs.CV↗

arXiv2016

Set-Theoretic Types for Polymorphic Variants

Polymorphic variants are a useful feature of the OCaml language whose current definition and implementation rely on kinding constraints to simulate a subtyping relation via unification. This yields an awkward formalization and results in a type system whose behaviour is in some cases unintuitive and/or unduly restrictive. In this work, we present an alternative formalization of poly-morphic variants, based on set-theoretic types and subtyping, that yields a cleaner and more streamlined system. Our formalization is more expressive than the current one (it types more programs while preserving type safety), it can internalize some meta-theoretic properties, and it removes some pathological cases of the current implementation resulting in a more intuitive and, thus, predictable type system. More generally, this work shows how to add full-fledged union types to functional languages of the ML family that usually rely on the Hindley-Milner type system. As an aside, our system also improves the theory of semantic subtyping, notably by proving completeness for the type reconstruction algorithm.

cs.PL↗

arXiv2011

Optimizing XML querying using type-based document projection

XML data projection (or pruning) is a natural optimization for main memory query engines: given a query Q over a document D, the subtrees of D that are not necessary to evaluate Q are pruned, thus producing a smaller document D'; the query Q is then executed on D', hence avoiding to allocate and process nodes that will never be reached by Q. In this article, we propose a new approach, based on types, that greatly improves current solutions. Besides providing comparable or greater precision and far lesser pruning overhead, our solution ---unlike current approaches--- takes into account backward axes, predicates, and can be applied to multiple queries rather than just to single ones. A side contribution is a new type system for XPath able to handle backward axes. The soundness of our approach is formally proved. Furthermore, we prove that the approach is also complete (i.e., yields the best possible type-driven pruning) for a relevant class of queries and Schemas. We further validate our approach using the XMark and XPathMark benchmarks and show that pruning not only improves the main memory query engine's performances (as expected) but also those of state of the art native XML databases.

cs.DB↗

arXiv2010

XPath Whole Query Optimization

Previous work reports about SXSI, a fast XPath engine which executes tree automata over compressed XML indexes. Here, reasons are investigated why SXSI is so fast. It is shown that tree automata can be used as a general framework for fine grained XML query optimization. We define the "relevant nodes" of a query as those nodes that a minimal automaton must touch in order to answer the query. This notion allows to skip many subtrees during execution, and, with the help of particular tree indexes, even allows to skip internal nodes of the tree. We efficiently approximate runs over relevant nodes by means of on-the-fly removal of alternation and non-determinism of (alternating) tree automata. We also introduce many implementation techniques which allows us to efficiently evaluate tree automata, even in the absence of special indexes. Through extensive experiments, we demonstrate the impact of the different optimization techniques.

cs.DB↗