Search Results for author:
Found 46 papers, 29 papers with code
GDumb: A Simple Approach that Questions Our Progress in Continual Learning
We discuss a general formulation for the Continual Learning (CL) problem for classification---a learning task where a stream provides samples to a learner and the goal of the learner, depending on the samples it receives, is to continually upgrade its knowledge about the old classes and learn new ones.
What Makes and Breaks Safety Fine-tuning? A Mechanistic Study
Safety fine-tuning helps align Large Language Models (LLMs) with human preferences for their safe deployment.
On Calibration of Object Detectors: Pitfalls, Evaluation and Baselines
To illustrate, D-DETR with our post-hoc Isotonic Regression calibrator outperforms the recent train-time state-of-the-art calibration method Cal-DETR by more than 7 D-ECE on the COCO dataset.
Placing Objects in Context via Inpainting for Out-of-distribution Segmentation
For example, we utilize it to augment Cityscapes samples by incorporating a subset of Pascal classes and demonstrate that models trained on such data achieve comparable performance to the Pascal-trained baseline.
Random Representations Outperform Online Continually Learned Representations
However, little attention has been paid to the efficacy of continually learned representations, as representations are learned alongside classifiers throughout the learning process.
Segment, Select, Correct: A Framework for Weakly-Supervised Referring Segmentation
Referring Image Segmentation (RIS) - the problem of identifying objects in images through natural language sentences - is a challenging task currently mostly solved through supervised learning.
MoCaE: Mixture of Calibrated Experts Significantly Improves Object Detection
Combining the strengths of many existing predictors to obtain a Mixture of Experts which is superior to its individual components is an effective way to improve the performance without having to develop new architectures or train a model from scratch.
Ranked #1 on
Object Detection In Aerial Images
on DOTA
(using extra training data)
Fine-tuning can cripple your foundation model; preserving features may be the solution
This is an undesirable effect of fine-tuning as a substantial amount of resources was used to learn these pre-trained concepts in the first place.
Towards Building Self-Aware Object Detectors via Reliable Uncertainty Quantification and Calibration
The current approach for testing the robustness of object detectors suffers from serious deficiencies such as improper methods of performing out-of-distribution detection and using calibration metrics which do not consider both localisation and classification quality.
Graph Inductive Biases in Transformers without Message Passing
Graph inductive biases are crucial for Graph Transformers, and previous works incorporate them using message-passing modules and/or positional encodings.
Ranked #1 on
Node Classification
on CLUSTER
Raising the Bar on the Evaluation of Out-of-Distribution Detection
In this paper, we define 2 categories of OoD data using the subtly different concepts of perceptual/visual and semantic similarity to in-distribution (iD) data.
Out-of-Distribution Detection
Out of Distribution (OOD) Detection
+2
Query-based Hard-Image Retrieval for Object Detection at Test Time
There is a longstanding interest in capturing the error behaviour of object detectors by finding images where their performance is likely to be unsatisfactory.
An Impartial Take to the CNN vs Transformer Robustness Contest
Following the surge of popularity of Transformers in Computer Vision, several studies have attempted to determine whether they could be more robust to distribution shifts and provide better uncertainty estimates than Convolutional Neural Networks (CNNs).
Sample-dependent Adaptive Temperature Scaling for Improved Calibration
The most common post-hoc approach to compensate for this is to perform temperature scaling, which adjusts the confidences of the predictions on any input by scaling the logits by a fixed value.
RegMixup: Mixup as a Regularizer Can Surprisingly Improve Accuracy and Out Distribution Robustness
We show that the effectiveness of the well celebrated Mixup [Zhang et al., 2018] can be further improved if instead of using it as the sole learning objective, it is utilized as an additional regularizer to the standard cross-entropy loss.
Catastrophic overfitting can be induced with discriminative non-robust features
Through extensive experiments we analyze this novel phenomenon and discover that the presence of these easy features induces a learning shortcut that leads to CO. Our findings provide new insights into the mechanisms of CO and improve our understanding of the dynamics of AT.
Make Some Noise: Reliable and Efficient Single-Step Adversarial Training
Recently, Wong et al. showed that adversarial training with single-step FGSM leads to a characteristic failure mode named Catastrophic Overfitting (CO), in which a model becomes suddenly vulnerable to multi-step attacks.
Mix-MaxEnt: Creating High Entropy Barriers To Improve Accuracy and Uncertainty Estimates of Deterministic Neural Networks
We propose an extremely simple approach to regularize a single deterministic neural network to obtain improved accuracy and reliable uncertainty estimates.
Towards fast and effective single-step adversarial training
In this work, we methodically revisit the role of noise and clipping in single-step adversarial training.
ANCER: Anisotropic Certification via Sample-wise Volume Maximization
Randomized smoothing has recently emerged as an effective tool that enables certification of deep neural network classifiers at scale.
No Cost Likelihood Manipulation at Test Time for Making Better Mistakes in Deep Networks
There has been increasing interest in building deep hierarchy-aware classifiers that aim to quantify and reduce the severity of mistakes, and not just reduce the number of errors.
How Benign is Benign Overfitting ?
We investigate two causes for adversarial vulnerability in deep neural networks: bad data and (poorly) trained models.
Amending Mistakes Post-hoc in Deep Networks by Leveraging Class Hierarchies
There has been increasing interest in building deep hierarchy-aware classifiers, aiming to quantify and reduce the severity of mistakes and not just count the number of errors.
On Batch Normalisation for Approximate Bayesian Inference
We study batch normalisation in the context of variational inference methods in Bayesian neural networks, such as mean-field or MC Dropout.
Continual Learning in Low-rank Orthogonal Subspaces
In continual learning (CL), a learner is faced with a sequence of tasks, arriving one after the other, and the goal is to remember all the tasks once the continual learning experience is finished.
Diagnosing and Preventing Instabilities in Recurrent Video Processing
Recurrent models are a popular choice for video enhancement tasks such as video denoising or super-resolution.
How benign is benign overfitting?
We investigate two causes for adversarial vulnerability in deep neural networks: bad data and (poorly) trained models.
Progressive Skeletonization: Trimming more fat from a network at initialization
Recent studies have shown that skeletonization (pruning parameters) of networks \textit{at initialization} provides all the practical benefits of sparsity both at inference and training time, while only marginally degrading their performance.
A Revised Generative Evaluation of Visual Dialogue
Evaluating Visual Dialogue, the task of answering a sequence of questions relating to a visual input, remains an open research challenge.
Calibrating Deep Neural Networks using Focal Loss
To facilitate the use of focal loss in practice, we also provide a principled approach to automatically select the hyperparameter involved in the loss function.
Using Hindsight to Anchor Past Knowledge in Continual Learning
In continual learning, the learner faces a stream of data whose distribution changes over time.
Mirror Descent View for Neural Network Quantization
Quantizing large Neural Networks (NN) while maintaining the performance is highly desirable for resource-limited devices due to reduced memory and time complexity.
Interactive Sketch & Fill: Multiclass Sketch-to-Image Translation
We propose an interactive GAN-based sketch-to-image translation method that helps novice users create images of simple objects.
Stable Rank Normalization for Improved Generalization in Neural Networks and GANs
Exciting new work on the generalization bounds for neural networks (NN) given by Neyshabur et al. , Bartlett et al. closely depend on two parameter-depenedent quantities: the Lipschitz constant upper-bound and the stable rank (a softer version of the rank operator).
Ranked #136 on
Image Generation
on CIFAR-10
On Tiny Episodic Memories in Continual Learning
But for a successful knowledge transfer, the learner needs to remember how to perform previous tasks.
Ranked #7 on
Class Incremental Learning
on cifar100
Visual Dialogue without Vision or Dialogue
We characterise some of the quirks and shortcomings in the exploration of Visual Dialogue - a sequential question-answering task where the questions and corresponding answers are related through given visual stimuli.
Proximal Mean-field for Neural Network Quantization
Compressing large Neural Networks (NN) by quantizing the parameters, while maintaining the performance is highly desirable due to reduced memory and time complexity.
Weakly-Supervised Learning of Metric Aggregations for Deformable Image Registration
Conventional approaches refer to the definition of a similarity criterion that, once endowed with a deformation model and a smoothness constraint, determines the optimal transformation to align two given images.
Robustness via Deep Low-Rank Representations
To achieve low dimensionality of learned representations, we propose an easy-to-use, end-to-end trainable, low-rank regularizer (LR) that can be applied to any intermediate layer representation of a DNN.
FlipDial: A Generative Model for Two-Way Visual Dialogue
We are the first to extend this paradigm to full two-way visual dialogue, where our model is capable of generating both questions and answers in sequence based on a visual input, for which we propose a set of novel evaluation measures and metrics.
Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigence
We observe that, in addition to forgetting, a known issue while preserving knowledge, IL also suffers from a problem we call intransigence, inability of a model to update its knowledge.
Deformable Registration through Learning of Context-Specific Metric Aggregation
We propose a novel weakly supervised discriminative algorithm for learning context specific registration metrics as a linear combination of conventional similarity measures.
Discovering Class-Specific Pixels for Weakly-Supervised Semantic Segmentation
We propose an approach to discover class-specific pixels for the weakly-supervised semantic segmentation task.
Weakly supervised Semantic Segmentation
Weakly-Supervised Semantic Segmentation
Multi-Agent Diverse Generative Adversarial Networks
Second, to enforce that different generators capture diverse high probability modes, the discriminator of MAD-GAN is designed such that along with finding the real and fake samples, it is also required to identify the generator that generated the given fake sample.
Minding the Gaps for Block Frank-Wolfe Optimization of Structured SVMs
In this paper, we propose several improvements on the block-coordinate Frank-Wolfe (BCFW) algorithm from Lacoste-Julien et al. (2013) recently used to optimize the structured support vector machine (SSVM) objective in the context of structured prediction, though it has wider applications.
Parsimonious Labeling
Furthermore, we propose an efficient graph-cuts based algorithm for the parsimonious labeling problem that provides strong theoretical guarantees on the quality of the solution.
