Siva Rajesh Kasa, Karan Gupta, Sumegh Roychowdhury, Ashutosh Kumar, Yaswanth Biruduraju, Santhosh Kumar Kasa, Pattisapu Nikhil Priyatam, Arindam Bhattacharya, Shailendra Agarwal, Vijay Huddar. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Semantic matching is an important component of a product search pipeline. Its goal is to capture the semantic intent of the search query as opposed to the syntactic matching performed by a lexical matching system. A semantic matching model captures relationships like synonyms, and also captures common behavioral patterns to retrieve relevant results by generalizing from purchase data. They however suffer from lack of availability of informative negative examples for model training. Various methods have been proposed in the past to address this issue based upon hard-negative mining and contrastive learning.

In many applications of Bloom filters, it is possible to exploit the patterns present in the inserted and non-inserted keys to achieve more compression than the standard Bloom filter. A new class of Bloom filters called Learned Bloom filters use machine learning models to exploit these patterns in the data. In practice, these methods and their variants raise many questions: the choice of machine learning models, the training paradigm to achieve the desired results, the choice of thresholds, the number of partitions in case multiple partitions are used, and other such design decisions. In this paper, we present a simple partitioned Bloom filter that works as follows: we partition the Bloom filter into segments, each of which uses a simple projection-based hash function computed using the data. We also provide a theoretical analysis that provides a principled way to select the design parameters of our method: number of hash functions and number of bits per partition. We perform empirical evaluations of our methods on various real-world datasets spanning several applications. We show that it can achieve an improvement in false positive rates of up to two orders of magnitude over standard Bloom filters for the same memory usage, and upto 50% better compression (bytes used per key) for same FPR, and, consistently beats the existing variants of learned Bloom filters.

Several applications, like malicious URL detection and web spam detection, require classification on very high-dimensional data. In such cases anomalous data is hard to find but normal data is easily available. As such it is increasingly common to use a one-class classifier (OCC). Unfortunately, most OCC algorithms cannot scale to datasets with extremely high dimensions. In this paper, we present Fast Random projection-based One-Class Classification (FROCC), an extremely efficient, scalable and easily parallelizable method for one-class classification with provable theoretical guarantees. Our method is based on the simple idea of transforming the training data by projecting it onto a set of random unit vectors that are chosen uniformly and independently from the unit sphere, and bounding the regions based on separation of the data. FROCC can be naturally extended with kernels. We provide a new theoretical framework to prove that that FROCC generalizes well in the sense that it is stable and has low bias for some parameter settings. We then develop a fast scalable approximation of FROCC using vectorization, exploiting data sparsity and parallelism to develop a new implementation called ParDFROCC. ParDFROCC achieves up to 2 percent points better ROC than the next best baseline, with up to 12× speedup in training and test times over a range of state-of-the-art benchmarks for the OCC task.

This work introduces a tool for interactive exploration and visualization using MetaTracts. MetaTracts is a novel method for extraction and visualization of individual fiber bundles and weaving patterns from X-ray computed tomography (XCT) scans of endless carbon fiber reinforced polymers (CFRPs). It is designed specifically to handle XCT scans of low resolutions where the individual fibers are barely visible, which makes extraction of fiber bundles a challenging problem. The proposed workflow is used to analyze unit cells of CFRP materials integrating a recurring weaving pattern. First, a coarse version of integral curves is used to trace sections of the individual fiber bundles in the woven CFRP materials. We call these sections MetaTracts. In the second step, these extracted fiber bundle sections are clustered using a two-step approach: first by orientation, then by proximity. The tool can generate volumetric representations as well as surface models of the extracted fiber bundles to be exported for further analysis. In addition a custom interactive tool for exploration and visual analysis of MetaTracts is designed. We evaluate the proposed workflow on a number of real world datasets and demonstrate that MetaTracts effectively and robustly identifies and extracts fiber bundles.

This work introduces MetaTracts, a novel method for extracting and visualizing individual fiber bundles and weaving patterns from X-ray computed tomography (XCT) scans of endless carbon fiber reinforced polymers (CFRP). The proposed work flow is designed to analyze unit cells of CFRP materials integrating the recurring weaving pattern. It is designed to handle XCT scans of low resolution, in which individual fibers are not visible or are barely visible. First, a coarse version of integral curves is used to trace subsections of the individual fiber bundles in the woven CFRP materials. We call these sections MetaTracts. In the second step, these extracted fiber bundle sections (MetaTracts) are clustered using a two-step approach: first by orientation, then by proximity. The tool can generate volumetric representations as well as surface models of the extracted fiber bundles to be exported for further analysis. We evaluate the proposed work flow on a number of real world datasets and demonstrate that MetaTracts effectively and robustly identifies and separates different fiber bundles.

Abstract A number of papers present algorithms to construct isosurfaces with sharp edges and corners from hermite data, i.e. the exact surface normals at the exact intersection of the surface and grid edges. We discuss some fundamental problems with the previous algorithms and describe a new approach, based on merging grid cubes near sharp edges, that produces significantly better results. Our algorithm requires only gradients at the grid vertices, not at each surface‐edge intersection point. We also give a method for measuring the correctness of the resulting sharp edges and corners in the isosurface.