Performance Degradation

Performance degradation in various machine learning and data processing systems is a significant research area focusing on identifying and mitigating factors that lead to reduced accuracy, efficiency, or reliability over time or under changing conditions. Current research investigates this across diverse architectures, including deep learning models (e.g., transformers, convolutional networks, graph convolutional networks), key-value stores, and algorithms for bandit problems, exploring issues like over-smoothing, gradient vanishing, and the impact of data distribution shifts. Understanding and addressing performance degradation is crucial for improving the robustness and reliability of AI systems in critical applications like healthcare, manufacturing, and data management, as well as enhancing the efficiency of fundamental data structures.

Papers

November 19, 2024

IMUVIE: Pickup Timeline Action Localization via Motion Movies
John Clapham, Kenneth Koltermann, Yanfu Zhang, Yuming Sun, Evie N Burnet, Gang Zhou
Fall Detection Action Localization Action Classification Performance Degradation Multi Agent Pickup Video Motion Wearable System

June 27, 2024

LearnedKV: Integrating LSM and Learned Index for Superior Performance on SSD
Wenlong Wang, David Hung-Chang Du
High Performance Learned Index Performance Degradation Neighbor Iou Voting SSD Detector Intelligent Reading Index Structure Merge Tree

June 1, 2024

A Batch Sequential Halving Algorithm without Performance Degradation
Sotetsu Koyamada, Soichiro Nishimori, Shin Ishii
Multi Armed Bandit Performance Degradation Batch Data

November 14, 2023

Are You Sure? Challenging LLMs Leads to Performance Drops in The FlipFlop Experiment
Philippe Laban, Lidiya Murakhovs'ka, Caiming Xiong, Chien-Sheng Wu
Large Language Model Open Source LLM Model Behavior Performance Degradation

August 16, 2023

It Ain't That Bad: Understanding the Mysterious Performance Drop in OOD Generalization for Generative Transformer Models
Xingcheng Xu, Zihao Pan, Haipeng Zhang, Yanqing Yang
Strong Generalization Transformer Based Model Improved Generalization Performance Degradation Generalization Behavior Generative Transformer Model

July 30, 2023

Towards Learned Predictability of Storage Systems
Chenyuan Wu
Client Availability Data Storage Performance Degradation Spatial Prediction Neighbor Iou Voting SSD Detector Storage System

July 23, 2023

Demonstration of a Response Time Based Remaining Useful Life (RUL) Prediction for Software Systems
Ray Islam, Peter Sandborn
Human Prediction Software System Useful Life Response Time Performance Degradation Prognostic and Health Management Open Source Software Reliability

June 21, 2023

Investigating Poor Performance Regions of Black Boxes: LIME-based Exploration in Sepsis Detection
Mozhgan Salimiparsa, Surajsinh Parmar, San Lee, Choongmin Kim, Yonghwan Kim, Jang Yong Kim
Inherent Interpretability Black Box Local Interpretable Model Agnostic Explanation Performance Degradation Agnostic Exploration Sepsis Detection Sepsis Diagnosis

June 8, 2023

RDumb: A simple approach that questions our progress in continual test-time adaptation
Ori Press, Steffen Schneider, Matthias Kümmerer, Matthias Bethge
Pre Trained Model Test Time Adaptation Much Progress Continual Test Time Adaptation Continual Adaptation Simple Approach Performance Degradation

May 25, 2023

Private Meeting Summarization Without Performance Loss
Seolhwa Lee, Anders Søgaard
Differential Privacy Unseen Entity Meeting Summarization Performance Degradation Differential Privacy Constraint

March 26, 2023

$\Delta$-Patching: A Framework for Rapid Adaptation of Pre-trained Convolutional Networks without Base Performance Loss
Chaitanya Devaguptapu, Samarth Sinha, K J Joseph, Vineeth N Balasubramanian, Animesh Garg
Zero Shot Transfer Learning New Framework Pre Trained Model Pre Trained Convolutional Neural Network Fast Adaptation Q Network Performance Degradation

November 9, 2022

Composite Fixed-Length Ordered Features for Palmprint Template Protection with Diminished Performance Loss
Weiqiang Zhao, Heng Zhao, Zhicheng Cao, Liaojun Pang
Feature Wise Performance Degradation Palmprint Recognition Variable Length

October 14, 2022

$\Lambda$-DARTS: Mitigating Performance Collapse by Harmonizing Operation Selection among Cells
Sajad Movahedi, Melika Adabinejad, Ayyoob Imani, Arezou Keshavarz, Mostafa Dehghani, Azadeh Shakery, Babak N. Araabi
Neural Architecture Search Gradient Based Optimization Differentiable Neural Architecture Search Inter Vendor Harmonization Optimal Architecture Performance Degradation

October 11, 2022

Performance Deterioration of Deep Learning Models after Clinical Deployment: A Case Study with Auto-segmentation for Definitive Prostate Cancer Radiotherapy
Biling Wang, Michael Dohopolski, Ti Bai, Junjie Wu, Raquibul Hannan, Neil Desai, Aurelie Garant, Daniel Yang, Dan Nguyen, Mu-Han Lin, Robert Timmerman, Xinlei Wang, Steve Jiang
Deep Learning Deep Learning Model Segmentation Model Prostate Cancer Automatic Segmentation Inadequate Contour Encoding Performance Degradation

August 10, 2022

A data-driven modular architecture with denoising autoencoders for health indicator construction in a manufacturing process
Emil Blixt Hansen, Helge Langseth, Nadeem Iftikhar, Simon Bøgh
Supervised Autoencoder Modular System Prognostic Prediction Modular Approach Performance Degradation Manufacturing Process Data Oriented Architecture

July 25, 2022

An Encryption Method of ConvMixer Models without Performance Degradation
Ryota Iijima, Hitoshi Kiya
Adversarial DEfense Privacy Preserving Image Encryption Method Performance Degradation ConvMixer Model

July 17, 2022

Performance degradation of ImageNet trained models by simple image transformations
Harsh Maheshwari
Full Model Supervised ImageNet Direct Convolution Pytorch Model Image Transformation Performance Degradation

December 22, 2021

SkipNode: On Alleviating Performance Degradation for Deep Graph Convolutional Networks
Weigang Lu, Yibing Zhan, Binbin Lin, Ziyu Guan, Liu Liu, Baosheng Yu, Wei Zhao, Yaming Yang, Dacheng Tao
Convolution Layer Undisciplined Over Smoothing Graph Convolutional Performance Degradation Smoothing Problem Skip List

Performance Degradation

Papers

IMUVIE: Pickup Timeline Action Localization via Motion Movies

LearnedKV: Integrating LSM and Learned Index for Superior Performance on SSD

A Batch Sequential Halving Algorithm without Performance Degradation

Are You Sure? Challenging LLMs Leads to Performance Drops in The FlipFlop Experiment

It Ain't That Bad: Understanding the Mysterious Performance Drop in OOD Generalization for Generative Transformer Models

Towards Learned Predictability of Storage Systems

Demonstration of a Response Time Based Remaining Useful Life (RUL) Prediction for Software Systems

Investigating Poor Performance Regions of Black Boxes: LIME-based Exploration in Sepsis Detection

RDumb: A simple approach that questions our progress in continual test-time adaptation

Private Meeting Summarization Without Performance Loss

$\Delta$-Patching: A Framework for Rapid Adaptation of Pre-trained Convolutional Networks without Base Performance Loss

Composite Fixed-Length Ordered Features for Palmprint Template Protection with Diminished Performance Loss

$\Lambda$-DARTS: Mitigating Performance Collapse by Harmonizing Operation Selection among Cells

Performance Deterioration of Deep Learning Models after Clinical Deployment: A Case Study with Auto-segmentation for Definitive Prostate Cancer Radiotherapy

A data-driven modular architecture with denoising autoencoders for health indicator construction in a manufacturing process

An Encryption Method of ConvMixer Models without Performance Degradation

Performance degradation of ImageNet trained models by simple image transformations

SkipNode: On Alleviating Performance Degradation for Deep Graph Convolutional Networks