Related Experiment Video
Updated: Jan 30, 2026
07:29
Online Explorative Study on the Learning Uses of Virtual Reality Among Early Adopters
Published on: November 22, 2019
8.6K
Exploring Frequencies via Feature Mixing and Meta-Learning for Improving Adversarial Transferability
Summary
This study introduces a frequency-based feature mixing method to improve adversarial attacks on Deep Neural Networks (DNNs). The approach enhances attack transferability against both normal and defense models by strategically combining features from clean and adversarial samples.
Area of Science:
- Artificial Intelligence
- Computer Vision
- Machine Learning Security
Background:
- Deep Neural Networks (DNNs) are vulnerable to adversarial attacks.
- Frequency-domain analysis reveals high-frequency components impact predictions, while low-frequency components aid black-box attack transferability.
Purpose of the Study:
- To develop a frequency decomposition-based feature mixing method to exploit frequency characteristics for improved adversarial attacks.
- To address the conflict arising from simultaneously applying different feature mixing strategies.
Main Methods:
- Introduced a frequency decomposition-based feature mixing method.
- Proposed a cross-frequency meta-optimization approach with meta-train, meta-test, and final update steps.
- Leveraged low-frequency components for defense model attacks and adversarial samples for normally-trained model attacks.
Main Results:
- Incorporating clean sample features into adversarial features is effective for normally-trained models.
- Combining clean features with low-frequency adversarial features improves attacks on defense models.
- The cross-frequency meta-optimization approach effectively enhances attack transferability against both model types.
Conclusions:
- The proposed frequency-based feature mixing and meta-optimization method significantly improves adversarial attack transferability.
- The study demonstrates a novel approach to exploit frequency characteristics for more effective DNN attacks.
- The method shows effectiveness on the ImageNet-Compatible dataset for both normally-trained and defense models.
Related Concept Videos
01:21Frequency-dependent Selection
23.9K
When the fitness of a trait is influenced by how common it is (i.e., its frequency) relative to different traits within a population, this is referred to as frequency-dependent selection. Frequency-dependent selection may occur between species or within a single species. This type of selection can either be positive—with more common phenotypes having higher fitness—or negative, with rarer phenotypes conferring increased fitness.
23.9K
01:09Directing Effect of Substituents: meta-Directing Groups
6.0K
Substituents on the benzene ring that direct an incoming electrophile to undergo substitution at the meta position are called meta directors. All meta directors either have a positive charge on the atom directly bonded to the ring or a partial positive charge. These groups function by withdrawing electrons from the ring through inductive and resonance effects. Consider the carbocation intermediates formed upon the addition of an electrophile on nitrobenzene at the...
6.0K
01:30Mixing Concrete
382
Concrete mixing ensures a homogenous blend where aggregates are well-coated with cement paste. Concrete mixing is typically done using two main types of mixers: batch and continuous. Batch mixers handle one batch at a time, thoroughly combining materials before discharging and receiving the next batch. In contrast, continuous mixers receive a steady flow of ingredients, mixing them consistently and discharging without interruption. Within batch mixers, tilting drum mixers mix with internal...
382
01:19Mixing Time
478
The concept of mixing time is significant in producing a uniform concrete mix with the required strength. The mixing period starts once all components are in the mixer. Initially, the mixer is charged with 10% of the water, followed by the consistent addition of solids and then 80% of the water. The remaining water is added later, within the first quarter of the mixing period. The minimum mixing time varies according to the mixer's capacity; for example, mixers with up to 1 cubic yard...
478
01:14Avoidance Learning and Learned Helplessness
2.6K
Avoidance learning and learned helplessness are critical concepts in understanding behavioral responses to negative stimuli.
Avoidance learning occurs when an organism learns that a specific behavior can prevent an unpleasant outcome. For example, a student who receives a bad grade may start studying harder to avoid future poor grades. This behavior persists even when the negative outcome is no longer present. Avoidance learning is powerful because it maintains behavior in the absence of the...
Avoidance learning occurs when an organism learns that a specific behavior can prevent an unpleasant outcome. For example, a student who receives a bad grade may start studying harder to avoid future poor grades. This behavior persists even when the negative outcome is no longer present. Avoidance learning is powerful because it maintains behavior in the absence of the...
2.6K
01:26Ready Mixed Concrete
365
Ready-mixed concrete, also known as pre-mixed concrete, is prepared in a centralized plant and then transported in trucks to construction sites where it is ready for placement. This type of concrete is categorized into central-mixed, truck-mixed (or transit-mixed), and shrink-mixed. Central-mixed concrete is entirely prepared at a plant and moved to the site in agitator trucks that rotate at a speed of 2 to 6 rpm. Truck-mixed concrete, on the other hand, has the ingredients batched at the plant...
365