Viral News

CoNO: Complex Neural Operator for Continous Dynamical Physical Systems

CoNO: Complex Neural Operator for Continous Dynamical Physical Systems

arXiv:2406.02597v1 Announce Type: new Abstract: Neural operators extend data-driven models to map between infinite-dimensional functional spaces. While these operators perform effectively in either the time or frequency domain, their performance may be limited when applied to non-stationary spatial or temporal signals whose frequency characteristics change with time. Here, we introduce Complex Neural Operator (CoNO) that parameterizes the integral kernel using Fractional Fourier Transform (FrFT), better representing non-stationary signals in a complex-valued domain. Theoretically, we prove the universal approximation capability of CoNO. We perform an extensive empirical evaluation of CoNO on seven challenging partial differential equations (PDEs), including regular grids, structured…
Read More
DenoDet: Attention as Deformable Multi-Subspace Feature Denoising for Target Detection in SAR Images

DenoDet: Attention as Deformable Multi-Subspace Feature Denoising for Target Detection in SAR Images

arXiv:2406.02833v1 Announce Type: new Abstract: Synthetic Aperture Radar (SAR) target detection has long been impeded by inherent speckle noise and the prevalence of diminutive, ambiguous targets. While deep neural networks have advanced SAR target detection, their intrinsic low-frequency bias and static post-training weights falter with coherent noise and preserving subtle details across heterogeneous terrains. Motivated by traditional SAR image denoising, we propose DenoDet, a network aided by explicit frequency domain transform to calibrate convolutional biases and pay more attention to high-frequencies, forming a natural multi-scale subspace representation to detect targets from the perspective of multi-subspace denoising. We design TransDeno, a…
Read More
Efficient Minimum Bayes Risk Decoding using Low-Rank Matrix Completion Algorithms

Efficient Minimum Bayes Risk Decoding using Low-Rank Matrix Completion Algorithms

arXiv:2406.02832v1 Announce Type: new Abstract: Minimum Bayes Risk (MBR) decoding is a powerful decoding strategy widely used for text generation tasks, but its quadratic computational complexity limits its practical application. This paper presents a novel approach for approximating MBR decoding using matrix completion techniques, focusing on the task of machine translation. We formulate MBR decoding as a matrix completion problem, where the utility metric scores between candidate hypotheses and pseudo-reference translations form a low-rank matrix. First, we empirically show that the scores matrices indeed have a low-rank structure. Then, we exploit this by only computing a random subset of the…
Read More
Opaque Launches New Platform For Running AI Workloads on Encrypted Data

Opaque Launches New Platform For Running AI Workloads on Encrypted Data

(JLStock/Shutterstock) While enterprises understand the need to innovate to stay competitive, they are also cautious about protecting their data. Enterprises often grapple with balancing innovation and security when it comes to extracting value from their data using generative AI. Existing techniques to operationalize the data are either too risky or inadequate. As a result, most organizations are forced to be cautious and prioritize security, resulting in stalled AI projects.  Opaque Systems, a security data analytics startup, offers a solution to overcome these challenges and unlock the full value of organizations’ data. The company has unveiled its new Confidential AI platform,…
Read More
DBRX at Data + AI Summit: Best Practices, Use Cases, and Behind-the-scenes

DBRX at Data + AI Summit: Best Practices, Use Cases, and Behind-the-scenes

Businesses are making remarkable progress on their data and AI journeys. They’re advancing from a few pilot projects confined to use cases likely deemed “non-mission-critical,” to deploying applications in real-world operations across thousands of users that are enhancing employee productivity and improving efficiency. However, building and scaling these next-generation systems is a complex, resource-intensive combination of engineering challenges up and down the stack. And as more users look to build their own GenAI applications, they’re encountering many of the same issues. At Databricks, we have learned quite a bit about those challenges firsthand. We used our Data Intelligence Platform to train our…
Read More
Slow and Steady Wins the Race: Maintaining Plasticity with Hare and Tortoise Networks

Slow and Steady Wins the Race: Maintaining Plasticity with Hare and Tortoise Networks

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. Source link lol
Read More
Distilling Aggregated Knowledge for Weakly-Supervised Video Anomaly Detection

Distilling Aggregated Knowledge for Weakly-Supervised Video Anomaly Detection

arXiv:2406.02831v1 Announce Type: new Abstract: Video anomaly detection aims to develop automated models capable of identifying abnormal events in surveillance videos. The benchmark setup for this task is extremely challenging due to: i) the limited size of the training sets, ii) weak supervision provided in terms of video-level labels, and iii) intrinsic class imbalance induced by the scarcity of abnormal events. In this work, we show that distilling knowledge from aggregated representations of multiple backbones into a relatively simple model achieves state-of-the-art performance. In particular, we develop a bi-level distillation approach along with a novel disentangled cross-attention-based feature aggregation network.…
Read More
Too Big to Fail: Larger Language Models are Disproportionately Resilient to Induction of Dementia-Related Linguistic Anomalies

Too Big to Fail: Larger Language Models are Disproportionately Resilient to Induction of Dementia-Related Linguistic Anomalies

arXiv:2406.02830v1 Announce Type: new Abstract: As artificial neural networks grow in complexity, understanding their inner workings becomes increasingly challenging, which is particularly important in healthcare applications. The intrinsic evaluation metrics of autoregressive neural language models (NLMs), perplexity (PPL), can reflect how "surprised" an NLM model is at novel input. PPL has been widely used to understand the behavior of NLMs. Previous findings show that changes in PPL when masking attention layers in pre-trained transformer-based NLMs reflect linguistic anomalies associated with Alzheimer's disease dementia. Building upon this, we explore a novel bidirectional attention head ablation method that exhibits properties attributed to…
Read More
What the Big Fuss Over Table Formats and Metadata Catalogs Is All About

What the Big Fuss Over Table Formats and Metadata Catalogs Is All About

(Pinare/Shutterstock) The big data community gained clarity on the future of data lakehouses earlier this week as a result of Snowflake’s open sourcing of its new Polaris metadata catalog and Databricks’ acquisition of Tabular. The actions cemented Apache Iceberg as the winner of the battle of open table formats, which is a big win for customers and open data, while it exposes a new competitive front: the metadata catalog. The news Monday and Tuesday was as hot as the weather in San Francisco this week, and left some longtime big data watchers gasping for breath. To recap: On Monday, Snowflake…
Read More
Azure Databricks at Databricks Data + AI Summit 2024 featuring Industry Leaders and Pioneers

Azure Databricks at Databricks Data + AI Summit 2024 featuring Industry Leaders and Pioneers

This is a collaborative post from Databricks and Microsoft. We thank Mohini Verma, Senior Product Marketing Manager, for her contributions.Data + AI Summit 2024: Register now to join this in-person and virtual event June 10-13 and learn from the global data community.Microsoft is a Legend Sponsor of the Databricks Data + AI Summit 2024, the premier event for the global data community. Join us to learn how data intelligence enables every organization to harness the power of generative AI on their own data. Hear from Microsoft leaders who will share how customers have successfully leveraged the Databricks Data Intelligence Platform…
Read More
No widgets found. Go to Widget page and add the widget in Offcanvas Sidebar Widget Area.