Viral News - CybAI news

21 Nov

Does Unlearning Truly Unlearn? A Black Box Evaluation of LLM Unlearning Methods

stp2y0 CommentsViral News

[Submitted on 18 Nov 2024 (v1), last revised 20 Nov 2024 (this version, v2)] View a PDF of the paper titled Does Unlearning Truly Unlearn? A Black Box Evaluation of LLM Unlearning Methods, by Jai Doshi and Asa Cooper Stickland View PDF HTML (experimental) Abstract:Large language model unlearning aims to remove harmful information that LLMs have learnt to prevent their use for malicious purposes. LLMU and RMU have been proposed as two methods for LLM unlearning, achieving impressive results on unlearning benchmarks. We study in detail the efficacy of these methods by evaluating their impact on general model capabilities on…

21 Nov

Select High-Level Features: Efficient Experts from a Hierarchical Classification Network

stp2y0 CommentsViral News

[Submitted on 8 Mar 2024 (v1), last revised 20 Nov 2024 (this version, v2)] View a PDF of the paper titled Select High-Level Features: Efficient Experts from a Hierarchical Classification Network, by Andr'e Kelm and 7 other authors View PDF HTML (experimental) Abstract:This study introduces a novel expert generation method that dynamically reduces task and computational complexity without compromising predictive performance. It is based on a new hierarchical classification network topology that combines sequential processing of generic low-level features with parallelism and nesting of high-level features. This structure allows for the innovative extraction technique: the ability to select only high-level…

21 Nov

DeSiRe-GS: 4D Street Gaussians for Static-Dynamic Decomposition and Surface Reconstruction for Urban Driving Scenes

stp2y0 CommentsViral News

arXiv:2411.11921v1 Announce Type: new Abstract: We present DeSiRe-GS, a self-supervised gaussian splatting representation, enabling effective static-dynamic decomposition and high-fidelity surface reconstruction in complex driving scenarios. Our approach employs a two-stage optimization pipeline of dynamic street Gaussians. In the first stage, we extract 2D motion masks based on the observation that 3D Gaussian Splatting inherently can reconstruct only the static regions in dynamic environments. These extracted 2D motion priors are then mapped into the Gaussian space in a differentiable manner, leveraging an efficient formulation of dynamic Gaussians in the second stage. Combined with the introduced geometric regularizations, our method are able…

21 Nov

HNCSE: Advancing Sentence Embeddings via Hybrid Contrastive Learning with Hard Negatives

stp2y0 CommentsViral News

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

21 Nov

KX Aims to Unlock Python’s Potential with PyKX 3.0

stp2y0 CommentsViral News

(dTosh/Shutterstock) As AI-driven algorithms become increasingly complex, the demand for scalable solutions that integrate powerful analytical engines with machine learning libraries has surged. KX, a performance analytical database for AI, has responded to this need by enhancing PyKX, its Python-first interface for kdb+, with a new hybrid architecture. PyKX 3.0 merges kdb+’s processing power with Python’s ML capabilities. The company claims that developers can use the platform to build advanced AI-driven applications and analytics without compromising on speed or scalability. In May 2023, KX had open-sourced PyKX, making its kdb+ time-series database and q programming language accessible to the global…

21 Nov

Refusal in LLMs is an Affine Function

stp2y0 CommentsViral News

[Submitted on 13 Nov 2024 (v1), last revised 19 Nov 2024 (this version, v2)] View a PDF of the paper titled Refusal in LLMs is an Affine Function, by Thomas Marshall and 2 other authors View PDF HTML (experimental) Abstract:We propose affine concept editing (ACE) as an approach for steering language models' behavior by intervening directly in activations. We begin with an affine decomposition of model activation vectors and show that prior methods for steering model behavior correspond to subsets of terms of this decomposition. We then provide a derivation of ACE and use it to control refusal behavior on…

21 Nov

Feasibility of Federated Learning from Client Databases with Different Brain Diseases and MRI Modalities

stp2y0 CommentsViral News

[Submitted on 17 Jun 2024 (v1), last revised 19 Nov 2024 (this version, v3)] View a PDF of the paper titled Feasibility of Federated Learning from Client Databases with Different Brain Diseases and MRI Modalities, by Felix Wagner and 8 other authors View PDF HTML (experimental) Abstract:Segmentation models for brain lesions in MRI are typically developed for a specific disease and trained on data with a predefined set of MRI modalities. Such models cannot segment the disease using data with a different set of MRI modalities, nor can they segment other types of diseases. Moreover, this training paradigm prevents a…

21 Nov

Retrieval-Augmented Personalization for Multimodal Large Language Models

stp2y0 CommentsViral News

[Submitted on 17 Oct 2024 (v1), last revised 18 Nov 2024 (this version, v2)] View a PDF of the paper titled Retrieval-Augmented Personalization for Multimodal Large Language Models, by Haoran Hao and 4 other authors View PDF HTML (experimental) Abstract:The development of large language models (LLMs) has significantly enhanced the capabilities of multimodal LLMs (MLLMs) as general assistants. However, lack of user-specific knowledge still restricts their application in human's daily life. In this paper, we introduce the Retrieval Augmented Personalization (RAP) framework for MLLMs' personalization. Starting from a general MLLM, we turn it into a personalized assistant in three steps.…

20 Nov

Four Steps to Go from Experimentation to Embedding AI Across the Enterprise

stp2y0 CommentsViral News

AI is everywhere. In just a couple of years, this technology has evolved significantly and is transforming the way most of us do business. And yet, many organizations continue to grapple with how they can really integrate AI into their daily operations. It’s critical that this changes soon. To thrive in the age of AI, companies must do more than simply adopt AI. They must embrace an iterative approach, continuously learning and adapting as the technology evolves. In this article, I’ll share four commitments that companies should make to transition to full AI adopters. Understand Your Business Challenges AI for…

20 Nov

Introducing Predictive Optimization for Statistics

stp2y0 CommentsViral News

We are excited to introduce the gated Public Preview of Predictive Optimization for statistics. Announced at the Data + AI Summit, Predictive Optimization is now generally available as an AI-driven approach to streamlining optimization processes. This feature currently supports essential data layout and cleanup tasks, and early feedback from users highlights its effectiveness in vastly simplifying routine data maintenance. With the addition of automatic statistics management, Predictive Optimization delivers customer value and simplifies operation through the following advancements:Intelligent selection of data-skipping statistics, eliminating the need for column order managementAutomatic collection of query optimization statistics, removing the necessity to run ANALYZE after…