Search

We propose a novel approach for humming transcription that combines a CNN-based architecture with a dynamic programming-based post-processing algorithm, utilizing the recently introduced HumTrans dataset. We identify and address inherent problems with the offset and onset ground truth provided by the dataset, offering heuristics to improve these annotations, resulting in a dataset with precise annotations that will aid future research. Additionally, we compare the transcription accuracy of our method against several others, demonstrating state-of-the-art (SOTA) results. All our code and corrected dataset is available at https://github.com/shubham-gupta-30/humming_transcription

This technical report introduces Docling, an easy to use, self-contained, MIT-licensed open-source package for PDF document conversion. It is powered by state-of-the-art specialized AI models for layout analysis (DocLayNet) and table structure recognition (TableFormer), and runs efficiently on commodity hardware in a small resource budget. The code interface allows for easy extensibility and addition of new features and models.

Environment, Social, and Governance (ESG) KPIs assess an organization's performance on issues such as climate change, greenhouse gas emissions, water consumption, waste management, human rights, diversity, and policies. ESG reports convey this valuable quantitative information through tables. Unfortunately, extracting this information is difficult due to high variability in the table structure as well as content. We propose Statements, a novel domain agnostic data structure for extracting quantitative facts and related information. We propose translating tables to statements as a new supervised deep-learning universal information extraction task. We introduce SemTabNet - a dataset of over 100K annotated tables. Investigating a family of T5-based Statement Extraction Models, our best model generates statements which are 82% similar to the ground-truth (compared to baseline of 21%). We demonstrate the advantages of statements by applying our model to over 2700 tables from ESG reports. The homogeneous nature of statements permits exploratory data analysis on expansive information found in large collections of ESG reports., Comment: Accepted at the NLP4Climate workshop in the 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024)

In this paper, we propose Phoneme Discretized Saliency Maps (PDSM), a discretization algorithm for saliency maps that takes advantage of phoneme boundaries for explainable detection of AI-generated voice. We experimentally show with two different Text-to-Speech systems (i.e., Tacotron2 and Fastspeech2) that the proposed algorithm produces saliency maps that result in more faithful explanations compared to standard posthoc explanation methods. Moreover, by associating the saliency maps to the phoneme representations, this methodology generates explanations that tend to be more understandable than standard saliency maps on magnitude spectrograms., Comment: Proc. Interspeech 2024, 3295-3299, doi: 10.21437/Interspeech.2024-632

With the success of machine learning (ML) applied to climate reaching further every day, emulators have begun to show promise not only for weather but for multi-year time scales in the atmosphere. Similar work for the ocean remains nascent, with state-of-the-art limited to models running for shorter time scales or only for regions of the globe. In this work, we demonstrate high-skill global emulation for surface ocean fields over 5-8 years of model rollout, accurately representing modes of variability for two different ML architectures (ConvNext and Transformers). In addition, we address the outstanding question of generalization, an essential consideration if the end-use of emulation is to model warming scenarios outside of the model training data. We show that 1) generalization is not an intrinsic feature of a data-driven emulator, 2) fine-tuning the emulator on only small amounts of additional data from a distribution similar to the test set can enable the emulator to perform well in a warmed climate, and 3) the forced emulators are robust to noise in the forcing.

In this thesis, we study problems at the interface of analysis and discrete mathematics. We discuss analogues of well known Hardy-type inequalities and Rearrangement inequalities on the lattice graphs $\mathbb{Z}^d$, with a particular focus on behaviour of sharp constants and optimizers.In the first half of the thesis, we analyse Hardy inequalities on $\mathbb{Z}^d$, first for $d=1$ and then for $d \geq 3$. We prove a sharp weighted Hardy inequality on integers with power weights of the form $n^\alpha$. This is done via two different methods, namely super-solution and Fourier method. We also use Fourier method to prove a weighted Hardy type inequality for higher order operators. After discussing the one dimensional case, we study the Hardy inequality in higher dimensions ($d \geq 3$). In particular, we compute the asymptotic behaviour of the sharp constant in the discrete Hardy inequality, as $d \rightarrow \infty$. This is done by converting the inequality into a continuous Hardy-type inequality on a torus for functions having zero average. These continuous inequalities are new and interesting in themselves. In the second half, we focus our attention on analogues of Rearrangement inequalities on lattice graphs. We begin by analysing the situation in dimension one. We define various notions of rearrangements and prove the corresponding Polya-Szeg\H{o} inequality. These inequalities are also applied to prove some weighted Hardy inequalities on integers. Finally, we study Rearrangement inequalities (Polya-Szeg\H{o}) on general graphs, with a particular focus on lattice graphs $\mathbb{Z}^d$, for $d \geq 2$. We develop a framework to study these inequalities, using which we derive concrete results in dimension two. In particular, these results develop connections between Polya-Szeg\H{o} inequality and various isoperimetric inequalities on graphs., Comment: PhD thesis, Imperial College London, 140 pages

Temporal graph neural networks Tgnn have exhibited state-of-art performance in future-link prediction tasks. Training of these TGNNs is enumerated by uniform random sampling based unsupervised loss. During training, in the context of a positive example, the loss is computed over uninformative negatives, which introduces redundancy and sub-optimal performance. In this paper, we propose modified unsupervised learning of Tgnn, by replacing the uniform negative sampling with importance-based negative sampling. We theoretically motivate and define the dynamically computed distribution for a sampling of negative examples. Finally, using empirical evaluations over three real-world datasets, we show that Tgnn trained using loss based on proposed negative sampling provides consistent superior performance., Comment: 10 pages

Pervasive use of social media has become the emerging source for real-time information (like images, text, or both) to identify various events. Despite the rapid growth of image and text-based event classification, the state-of-the-art (SOTA) models find it challenging to bridge the semantic gap between features of image and text modalities due to inconsistent encoding. Also, the black-box nature of models fails to explain the model's outcomes for building trust in high-stakes situations such as disasters, pandemic. Additionally, the word limit imposed on social media posts can potentially introduce bias towards specific events. To address these issues, we proposed CrisisKAN, a novel Knowledge-infused and Explainable Multimodal Attention Network that entails images and texts in conjunction with external knowledge from Wikipedia to classify crisis events. To enrich the context-specific understanding of textual information, we integrated Wikipedia knowledge using proposed wiki extraction algorithm. Along with this, a guided cross-attention module is implemented to fill the semantic gap in integrating visual and textual data. In order to ensure reliability, we employ a model-specific approach called Gradient-weighted Class Activation Mapping (Grad-CAM) that provides a robust explanation of the predictions of the proposed model. The comprehensive experiments conducted on the CrisisMMD dataset yield in-depth analysis across various crisis-specific tasks and settings. As a result, CrisisKAN outperforms existing SOTA methodologies and provides a novel view in the domain of explainable multimodal event classification.

In this paper, we present a comprehensive evaluation to establish a robust and efficient framework for Lagrangian-based particle tracing using deep neural networks (DNNs). Han et al. (2021) first proposed a DNN-based approach to learn Lagrangian representations and demonstrated accurate particle tracing for an analytic 2D flow field. In this paper, we extend and build upon this prior work in significant ways. First, we evaluate the performance of DNN models to accurately trace particles in various settings, including 2D and 3D time-varying flow fields, flow fields from multiple applications, flow fields with varying complexity, as well as structured and unstructured input data. Second, we conduct an empirical study to inform best practices with respect to particle tracing model architectures, activation functions, and training data structures. Third, we conduct a comparative evaluation of prior techniques that employ flow maps as input for exploratory flow visualization. Specifically, we compare our extended model against its predecessor by Han et al. (2021), as well as the conventional approach that uses triangulation and Barycentric coordinate interpolation. Finally, we consider the integration and adaptation of our particle tracing model with different viewers. We provide an interactive web-based visualization interface by leveraging the efficiencies of our framework, and perform high-fidelity interactive visualization by integrating it with an OSPRay-based viewer. Overall, our experiments demonstrate that using a trained DNN model to predict new particle trajectories requires a low memory footprint and results in rapid inference. Following best practices for large 3D datasets, our deep learning approach using GPUs for inference is shown to require approximately 46 times less memory while being more than 400 times faster than the conventional methods., Comment: Accepted by Pacific Vis 2024

Despite the great success of Neural Radiance Fields (NeRF), its data-gathering process remains vague with only a general rule of thumb of sampling as densely as possible. The lack of understanding of what actually constitutes good views for NeRF makes it difficult to actively plan a sequence of views that yield the maximal reconstruction quality. We propose Surrogate Objectives for Active Radiance Fields (SOAR), which is a set of interpretable functions that evaluates the goodness of views using geometric and photometric visual cues - surface coverage, geometric complexity, textural complexity, and ray diversity. Moreover, by learning to infer the SOAR scores from a deep network, SOARNet, we are able to effectively select views in mere seconds instead of hours, without the need for prior visits to all the candidate views or training any radiance field during such planning. Our experiments show SOARNet outperforms the baselines with $\sim$80x speed-up while achieving better or comparable reconstruction qualities. We finally show that SOAR is model-agnostic, thus it generalizes across fully neural-implicit to fully explicit approaches., Comment: 13 pages

Deep graph generative modeling has gained enormous attraction in recent years due to its impressive ability to directly learn the underlying hidden graph distribution. Despite their initial success, these techniques, like much of the existing deep generative methods, require a large number of training samples to learn a good model. Unfortunately, large number of training samples may not always be available in scenarios such as drug discovery for rare diseases. At the same time, recent advances in few-shot learning have opened door to applications where available training data is limited. In this work, we introduce the hitherto unexplored paradigm of few-shot graph generative modeling. Towards this, we develop GSHOT, a meta-learning based framework for few-shot labeled graph generative modeling. GSHOT learns to transfer meta-knowledge from similar auxiliary graph datasets. Utilizing these prior experiences, GSHOT quickly adapts to an unseen graph dataset through self-paced fine-tuning. Through extensive experiments on datasets from diverse domains having limited training samples, we establish that GSHOT generates graphs of superior fidelity compared to existing baselines., Comment: Accepted in Learning on Graph Conference (LOG,2023),https://openreview.net/forum?id=Hy9K2WiVwW

Graph neural networks (GNNs), in general, are built on the assumption of a static set of features characterizing each node in a graph. This assumption is often violated in practice. Existing methods partly address this issue through feature imputation. However, these techniques (i) assume uniformity of feature set across nodes, (ii) are transductive by nature, and (iii) fail to work when features are added or removed over time. In this work, we address these limitations through a novel GNN framework called GRAFENNE. GRAFENNE performs a novel allotropic transformation on the original graph, wherein the nodes and features are decoupled through a bipartite encoding. Through a carefully chosen message passing framework on the allotropic transformation, we make the model parameter size independent of the number of features and thereby inductive to both unseen nodes and features. We prove that GRAFENNE is at least as expressive as any of the existing message-passing GNNs in terms of Weisfeiler-Leman tests, and therefore, the additional inductivity to unseen features does not come at the cost of expressivity. In addition, as demonstrated over four real-world graphs, GRAFENNE empowers the underlying GNN with high empirical efficacy and the ability to learn in continual fashion over streaming feature sets., Comment: 17 pages, 4 figures and 9 tables. Accepted in ICML 2023, DOI will be updated once it is available

We study Hardy inequalities for antisymmetric functions in three different settings: euclidean space, torus and the integer lattice. In particular, we show that under the antisymmetric condition the sharp constant in Hardy inequality increases substantially and grows as d^4 as d \rightarrow \infty in all cases. As a side product, we prove Hardy inequality on a domain whose boundary forms a corner at the point of singularity x=0., Comment: Minor changes

This research study investigates the efficiency of different information retrieval (IR) systems in accessing relevant information from the scientific literature during the COVID-19 pandemic. The study applies the TREC framework to the COVID-19 Open Research Dataset (CORD-19) and evaluates BM25, Contriever, and Bag of Embeddings IR frameworks. The objective is to build a test collection for search engines that tackle the complex information landscape during a pandemic. The study uses the CORD-19 dataset to train and evaluate the IR models and compares the results to those manually labeled in the TREC-COVID IR Challenge. The results indicate that advanced IR models like BERT and Contriever better retrieve relevant information during a pandemic. However, the study also highlights the challenges in processing large datasets and the need for strategies to focus on abstracts or summaries. Overall, the research highlights the importance of effectively tailored IR systems in dealing with information overload during crises like COVID-19 and can guide future research and development in this field., Comment: 7 pages, 2 figures