paper-with-me

홈 › Papers

Simple Techniques Work Surprisingly Well for Neural Network Test Prioritization and Active Learning (Replicability Study)

2022-05-02 · Michael Weiss, Paolo Tonella

Test Input Prioritizers (TIP) for Deep Neural Networks (DNN) are an important technique to handle the typically very large test datasets efficiently, saving computation and labeling costs. This is particularly true for large-scale, deployed systems, where inputs observed in production are recorded to serve as potential test or training data for the next versions of the system. Feng et. al. propose DeepGini, a very fast and simple TIP, and show that it outperforms more elaborate techniques such as neuron- and surprise coverage. In a large-scale study (4 case studies, 8 test datasets, 32'200 trained models) we verify their findings. However, we also find that other comparable or even simpler baselines from the field of uncertainty quantification, such as the predicted softmax likelihood or the entropy of the predicted softmax likelihoods perform equally well as DeepGini.

📄 PDF Abstract BibTeX arXiv:2205.00664

Code (3)

testingautomated-usi/corrupted-text 공식 구현
testingautomated-usi/dnn-tip 공식 구현
testingautomated-usi/simple-tip 공식 구현 tf

Tasks

Active LearningUncertainty Quantification

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Similar Papers 제목 키워드 기반

Testing the Limits of Jailbreaking Defenses with the Purple Problem

2024-03-20 · Taeyoun Kim, Suhas Kotha, aditi raghunathan

The rise of "jailbreak" attacks on language models has led to a flurry of defenses aimed at preventing undesirable responses. We critically examine the two stages of the defense pipeline: (i) defining what constitutes un…

Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability

2021-07-13 · NeurIPS 2021 12 · Dibya Ghosh, Jad Rahme, Aviral Kumar, Amy Zhang 외

Generalization is a central challenge for the deployment of reinforcement learning (RL) systems in the real world. In this paper, we show that the sequential structure of the RL problem necessitates new approaches to gen…

Reinforcement Learning (RL)

A Frustratingly Simple Decoding Method for Neural Text Generation

2023-05-22 · Haoran Yang, Deng Cai, Huayang Li, Wei Bi 외

We introduce a frustratingly simple, super efficient and surprisingly effective decoding method, which we call Frustratingly Simple Decoding (FSD), for neural text generation. The idea behind FSD is straightforward: we b…

Language ModelingLanguage ModellingText Generation

Freeze the Discriminator: a Simple Baseline for Fine-Tuning GANs

2020-02-25 · Sangwoo Mo, Minsu Cho, Jinwoo Shin

Generative adversarial networks (GANs) have shown outstanding performance on a wide range of problems in computer vision, graphics, and machine learning, but often require numerous training data and heavy computational r…

10-shot image generationImage GenerationTransfer Learning

Rethinking Complex Neural Network Architectures for Document Classification

2019-06-01 · NAACL 2019 6 · Ashutosh Adhikari, Achyudh Ram, Raphael Tang, Jimmy Lin

Neural network models for many NLP tasks have grown increasingly complex in recent years, making training and deployment more difficult. A number of recent papers have questioned the necessity of such architectures and f…

ClassificationDocument ClassificationGeneral ClassificationLanguage Modeling+1