Search Results for author:
Found 60 papers, 34 papers with code
Dataset Size Recovery from LoRA Weights
In this paper, we introduce a new task: dataset size recovery, that aims to determine the number of samples used to train a model, directly from its weights.
Real-Time Deepfake Detection in the Real-World
The image deepfake score is the pooled score of its patches.
From Zero to Hero: Cold-Start Anomaly Detection
This paper studies the realistic but underexplored cold-start setting where an anomaly detection model is initialized using zero-shot guidance, but subsequently receives a small number of contaminated observations (namely, that may include anomalies).
Ranked #1 on
Cold-Start Anomaly Detection
on BANKING77-OOS
On the Origin of Llamas: Model Tree Heritage Recovery
However, this information is underutilized as the weights are uninterpretable, and publicly available models are disorganized.
ObjectDrop: Bootstrapping Counterfactuals for Photorealistic Object Removal and Insertion
To tackle this challenge, we propose bootstrap supervision; leveraging our object removal model trained on a small counterfactual dataset, we synthetically expand this dataset considerably.
Distilling Datasets Into Less Than One Image
Current methods frame this as maximizing the distilled classification accuracy for a budget of K distilled images-per-class, where K is a positive integer.
Ranked #1 on
Dataset Distillation - 1IPC
on CUB-200-2011
Recovering the Pre-Fine-Tuning Weights of Generative Models
The dominant paradigm in generative modeling consists of two steps: i) pre-training on a large-scale but unsafe dataset, ii) aligning the pre-trained model with human values via fine-tuning.
Classifying Nodes in Graphs without GNNs
Recently, distillation methods succeeded in eliminating the use of GNNs at test time but they still require them during training.
Set Features for Anomaly Detection
This paper proposes to use set features for detecting anomalies in samples that consist of unusual combinations of normal elements.
Ranked #4 on
Anomaly Detection
on MVTec LOCO AD
Detecting Deepfakes Without Seeing Any
We therefore introduce the concept of "fact checking", adapted from fake news detection, for detecting zero-day deepfake attacks.
Ranked #1 on
DeepFake Detection
on FakeAVCeleb
Representation Learning in Anomaly Detection: Successes, Limits and a Grand Challenge
These limitations can be overcome when there are strong tasks priors, as is the case for many industrial tasks.
No Free Lunch: The Hazards of Over-Expressive Representations in Anomaly Detection
It is tempting to hypothesize that anomaly detection can improve indefinitely by increasing the scale of our networks, making their representations more expressive.
Unsupervised Word Segmentation Using Temporal Gradient Pseudo-Labels
We use a thresholding function on the temporal gradient magnitude to define a psuedo-label for wordness.
Set Features for Fine-grained Anomaly Detection
Fine-grained anomaly detection has recently been dominated by segmentation based approaches.
Ranked #1 on
Anomaly Detection
on UEA time-series datasets
Dreamix: Video Diffusion Models are General Video Editors
Our approach uses a video diffusion model to combine, at inference time, the low-resolution spatio-temporal information from the original video with new, high resolution information that it synthesized to align with the guiding text prompt.
Improving Zero-Shot Models with Label Distribution Priors
We propose a new approach, CLIPPR (CLIP with Priors), which adapts zero-shot models for regression and classification on unlabelled datasets.
Attribute-based Representations for Accurate and Interpretable Video Anomaly Detection
Surprisingly, we find that this simple representation is sufficient to achieve state-of-the-art performance in ShanghaiTech, the largest and most complex VAD dataset.
Ranked #1 on
Abnormal Event Detection In Video
on UCSD Ped2
Conffusion: Confidence Intervals for Diffusion Models
Diffusion models have become the go-to method for many generative tasks, particularly for image-to-image generation tasks such as super-resolution and inpainting.
Anomaly Detection Requires Better Representations
Anomaly detection seeks to identify unusual phenomena, a central task in science and industry.
Ranked #1 on
Anomaly Detection
on ODDS
Red PANDA: Disambiguating Anomaly Detection by Removing Nuisance Factors
Breaking from previous research, we present a new anomaly detection method that allows operators to exclude an attribute from being considered as relevant for anomaly detection.
A Contrastive Objective for Learning Disentangled Representations
Here, our objective is to learn representations that are invariant to the domain (sensitive attribute) for which labels are provided, while being informative over all other image attributes, which are unlabeled.
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection
We utilize a recently introduced 3D anomaly detection dataset to evaluate whether or not using 3D information is a lost opportunity.
Ranked #3 on
3D Anomaly Detection and Segmentation
on MVTEC 3D-AD
3D Anomaly Detection
3D Anomaly Detection and Segmentation
+3
Time Series Anomaly Detection by Cumulative Radon Features
We show that by parameterizing each time series using cumulative Radon features, we are able to efficiently and effectively model the distribution of normal time series.
Out-of-Distribution Detection Without Class Labels
Out-of-distribution detection seeks to identify novelties, samples that deviate from the norm.
Ranked #1 on
Anomaly Detection
on Anomaly Detection on Anomaly Detection on Unlabeled ImageNet-30 vs Flowers-102
(using extra training data)
Semantic Segmentation In-the-Wild Without Seeing Any Segmentation Examples
The output of this stage provides pixel-level pseudo-labels, instead of the manual pixel-level labels required by supervised methods.
The Inductive Bias of In-Context Learning: Rethinking Pretraining Example Design
We highlight a bias introduced by this common practice: we prove that the pretrained NLM can model much stronger dependencies between text segments that appeared in the same training example, than it can between text segments that appeared in different training examples.
Inductive-Biases for Contrastive Learning of Disentangled Representations
Current discriminative approaches are typically based on adversarial-training and do not reach comparable accuracy.
Fieldwise Factorized Networks for Tabular Data Classification
Tabular data is one of the most common data-types in machine learning, however, deep neural networks have not yet convincingly outperformed classical baselines on such datasets.
Language-Guided Image Clustering
In this setting, the model is provided with an exhaustive list of phrases describing all the possible values of a specific attribute, together with a shared image-language embedding (e. g.
Image Shape Manipulation from a Single Augmented Training Sample
In this paper, we present DeepSIM, a generative model for conditional image manipulation based on a single image.
Ranked #1 on
Image Manipulation
on LRS2
An Image is Worth More Than a Thousand Words: Towards Disentanglement in the Wild
Unsupervised disentanglement has been shown to be theoretically impossible without inductive biases on the models and the data.
Mean-Shifted Contrastive Loss for Anomaly Detection
We take the approach of transferring representations pre-trained on external datasets for anomaly detection.
Ranked #4 on
Anomaly Detection
on One-class CIFAR-100
(using extra training data)
Dataset Summarization by K Principal Concepts
The output of our method is a set of K principal concepts that summarize the dataset.
Ranked #4 on
Image Clustering
on ImageNet-100
(using extra training data)
Scaling-up Disentanglement for Image Translation
In this work, we propose OverLORD, a single framework for disentangling labeled and unlabeled attributes as well as synthesizing high-fidelity images, which is composed of two stages; (i) Disentanglement: Learning disentangled representations with latent optimization.
Membership Inference Attacks are Easier on Difficult Problems
Membership inference attacks (MIA) try to detect if data samples were used to train a neural network model, e. g. to detect copyright abuses.
Learning Disentangled Representations for Image Translation
Recent approaches for unsupervised image translation are strongly reliant on generative adversarial training and architectural locality constraints.
Membership Attacks on Conditional Generative Models Using Image Difficulty
A simple but effective approach for membership attacks can therefore use the reconstruction error.
PANDA: Adapting Pretrained Features for Anomaly Detection and Segmentation
In recent years, the anomaly detection community has attempted to obtain better features using advances in deep self-supervised feature learning.
Ranked #1 on
Anomaly Detection
on Cats-and-Dogs
Improving Style-Content Disentanglement in Image-to-Image Translation
Unsupervised image-to-image translation methods have achieved tremendous success in recent years.
Image Shape Manipulation from a Single Augmented Training Sample
In this paper, we present DeepSIM, a generative model for conditional image manipulation based on a single image.
Sub-Image Anomaly Detection with Deep Pyramid Correspondences
Nearest neighbor (kNN) methods utilizing deep pre-trained features exhibit very strong anomaly detection performance when applied to entire images.
Ranked #23 on
Anomaly Detection
on VisA
Classification-Based Anomaly Detection for General Data
Anomaly detection, finding patterns that substantially deviate from those seen previously, is one of the fundamental problems of artificial intelligence.
Ranked #2 on
Anomaly Detection
on UEA time-series datasets
Training End-to-end Single Image Generators without GANs
We present AugurOne, a novel approach for training single image generative models.
Deep Nearest Neighbor Anomaly Detection
Nearest neighbors is a successful and long-standing technique for anomaly detection.
Ranked #2 on
Anomaly Detection
on Anomaly Detection on Unlabeled ImageNet-30 vs CUB-200
(using extra training data)
Crypto-Oriented Neural Architecture Design
As neural networks revolutionize many applications, significant privacy conflicts between model users and providers emerge.
Demystifying Inter-Class Disentanglement
Learning to disentangle the hidden factors of variations within a set of observations is a key task for artificial intelligence.
Style Generator Inversion for Image Enhancement and Animation
We show that style generators outperform other GANs as well as Deep Image Prior as priors for image enhancement tasks.
Non-Adversarial Image Synthesis with Generative Latent Nearest Neighbors
GLANN combines the strengths of IMLE and GLO in a way that overcomes the main drawbacks of each method.
Towards Unsupervised Single-Channel Blind Source Separation using Adversarial Pair Unmix-and-Remix
Blind single-channel source separation is a long standing signal processing challenge.
Non-Adversarial Mapping with VAEs
Our method is much faster at inference time, is able to leverage large datasets and has a simple interpretation.
Neural separation of observed and unobserved distributions
In this work, we introduce a new method---Neural Egg Separation---to tackle the scenario of extracting a signal from an unobserved distribution additively mixed with a signal from an observed distribution.
Neural Separation of Observed and Unobserved Distribution
In this work, we introduce a new method---Neural Egg Separation---to tackle the scenario of extracting a signal from an unobserved distribution additively mixed with a signal from an observed distribution.
NAM: Non-Adversarial Unsupervised Domain Mapping
NAM relies on a pre-trained generative model of the target domain, and aligns each source image with an image synthesized from the target domain, while jointly optimizing the domain mapping function.
Unsupervised Correlation Analysis
Linking between two data sources is a basic building block in numerous computer vision problems.
Non-Adversarial Unsupervised Word Translation
We present a novel method that first aligns the second moment of the word distributions of the two languages and then iteratively refines the alignment.
Identifying Analogies Across Domains
We further show that the cross-domain mapping task can be broken into two parts: domain alignment and learning the mapping function.
VAIN: Attentional Multi-agent Predictive Modeling
In this paper we introduce VAIN, a novel attentional architecture for multi-agent predictive modeling that scales linearly with the number of agents.
Visual Learning of Arithmetic Operations
This indicates that while some tasks may be easily learnable end-to-end, other may need to be broken into sub-tasks.
Live Video Synopsis for Multiple Cameras
Video surveillance cameras generate most of recorded video, and there is far more recorded video than operators can watch.
An Egocentric Look at Video Photographer Identity
As head-worn cameras do not capture the photographer, it may seem that the anonymity of the photographer is preserved even when the video is publicly distributed.
