565 citations · 1.1k across the 18 of their papers we have counts for
19 papers
TAB: Transformer Attention Bottlenecks enable User Intervention and Debugging in Vision-Language Models
Pooyan Rahmanzadehgervi, Hung Huy Nguyen, Rosanne Liu +2
Multi-head self-attention (MHSA) is a key component of Transformers, a widely popular architecture in both language and vision. Multiple heads intuitively enable different parallel…
Logit Scaling for Out-of-Distribution Detection
Andrija Djurisic, Rosanne Liu, Mladen Nikolic
The safe deployment of machine learning and AI models in open-world settings hinges critically on the ability to detect out-of-distribution (OOD) data accurately, data samples that…
Character-Aware Models Improve Visual Text Rendering
Rosanne Liu, Dan Garrette, Chitwan Saharia +7
Current image generation models struggle to reliably produce well-formed visual text. In this paper, we investigate a key contributing factor: popular text-to-image models lack cha…
Extremely Simple Activation Shaping for Out-of-Distribution Detection
Andrija Djurisic, Nebojsa Bozanic, Arjun Ashok +1
The separation between training and deployment of machine learning models implies that not all scenarios encountered in deployment can be anticipated during training, and therefore…
What does a platypus look like? Generating customized prompts for zero-shot image classification
Sarah Pratt, Ian Covert, Rosanne Liu +1
Open-vocabulary models are a promising new paradigm for image classification. Unlike traditional classification models, open-vocabulary models classify among any arbitrary set of c…
Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models
Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao +448
Language models demonstrate both quantitative improvement and new qualitative capabilities with increasing scale. Despite their potentially transformative impact, these new capabil…
When less is more: Simplifying inputs aids neural network understanding
Robin Tibor Schirrmeister, Rosanne Liu, Sara Hooker +1
How do neural network image classifiers respond to simpler and simpler inputs? And what do such responses reveal about the learning process? To answer these questions, we need a cl…
Natural Adversarial Objects
Felix Lau, Nishant Subramani, Sasha Harrison +3
Although state-of-the-art object detection methods have shown compelling performance, models often are not robust to adversarial attacks and out-of-distribution data. We introduce…
Language Models are Few-shot Multilingual Learners
Genta Indra Winata, Andrea Madotto, Zhaojiang Lin +3
General-purpose language models have demonstrated impressive capabilities, performing on par with state-of-the-art approaches on a range of downstream natural language processing (…
When does loss-based prioritization fail?
Niel Teng Hu, Xinyu Hu, Rosanne Liu +2
Not all examples are created equal, but standard deep neural network training protocols treat each training point uniformly. Each example is propagated forward and backward through…
Why is Pruning at Initialization Immune to Reinitializing and Shuffling?
Sahib Singh, Rosanne Liu
Recent studies assessing the efficacy of pruning neural networks methods uncovered a surprising finding: when conducting ablation studies on existing pruning-at-initialization meth…
Supermasks in Superposition
Mitchell Wortsman, Vivek Ramanujan, Rosanne Liu +4
We present the Supermasks in Superposition (SupSup) model, capable of sequentially learning thousands of tasks without catastrophic forgetting. Our approach uses a randomly initial…
Estimating Q(s,s') with Deep Deterministic Dynamics Gradients
Ashley D. Edwards, Himanshu Sahni, Rosanne Liu +7
In this paper, we introduce a novel form of value function, , that expresses the utility of transitioning from a state to a neighboring state and then acting opt…
Plug and Play Language Models: A Simple Approach to Controlled Text Generation
Sumanth Dathathri, Andrea Madotto, Janice Lan +5
Large transformer-based language models (LMs) trained on huge text corpora have shown unparalleled generation capabilities. However, controlling attributes of the generated languag…
LCA: Loss Change Allocation for Neural Network Training
Janice Lan, Rosanne Liu, Hattie Zhou +1
Neural networks enjoy widespread use, but many aspects of their training, representation, and operation are poorly understood. In particular, our view into the training process is…
Deconstructing Lottery Tickets: Zeros, Signs, and the Supermask
Hattie Zhou, Janice Lan, Rosanne Liu +1
The recent "Lottery Ticket Hypothesis" paper by Frankle & Carbin showed that a simple approach to creating sparse networks (keeping the large weights) results in models that are tr…
An Atari Model Zoo for Analyzing, Visualizing, and Comparing Deep Reinforcement Learning Agents
Felipe Petroski Such, Vashisht Madhavan, Rosanne Liu +8
Much human and computational effort has aimed to improve how deep reinforcement learning algorithms perform on benchmarks such as the Atari Learning Environment. Comparatively less…
An Intriguing Failing of Convolutional Neural Networks and the CoordConv Solution
Rosanne Liu, Joel Lehman, Piero Molino +4
Few ideas have enjoyed as large an impact on deep learning as convolution. For any problem involving pixels or spatial representations, common intuition holds that convolutional ne…
Measuring the Intrinsic Dimension of Objective Landscapes
Chunyuan Li, Heerad Farkhoor, Rosanne Liu +1
Many recently trained neural networks employ large numbers of parameters to achieve good performance. One may intuitively use the number of parameters required as a rough gauge of…