ETT: Expanding the Long Context Understanding Capability of LLMs at Test-Time

Transformer-based Language Models' computation and memory overhead increase quadratically as a function of sequence length. The quadratic cost poses challenges when employing LLMs for processing long sequences. In this work, we introduce \ourmodelacronym~(Extend at Test-Time), method for extending the context length of short context Transformer-based LLMs, with constant memory requirement and linear computation overhead. ETT enable the extension of the context length at test-time by efficient fine-tuning the model's parameters on the input context, chunked into overlapping small subsequences. We evaluate ETT on LongBench by extending the context length of GPT-Large and Phi-2 up to 32 times, increasing from 1k to 32k tokens. This results in up to a 30 percent improvement in the model's accuracy. We also study how context can be stored in LLM's weights effectively and efficiently. Through a detailed ablation study, we examine which Transformer modules are most beneficial to fine-tune at test-time. Interestingly, we find that fine-tuning the second layer of the FFNs is more effective than full fine-tuning, leading to a further improvement in the models' accuracy.

Clarifying Myths About the Relationship Between Shape Bias, Accuracy, and Robustness

Deep learning models can perform well when evaluated on images from the same distribution as the training set. However, applying small perturbations in the forms of noise, artifacts, occlusions, blurring, etc. to a model's input image and feeding the model with out-of-distribution (OOD) data can significantly drop the model's accuracy, making it not applicable to real-world scenarios. Data augmentation is one of the well-practiced methods to improve model robustness against OOD data; however, examining which augmentation type to choose and how it affects the OOD robustness remains understudied. There is a growing belief that augmenting datasets using data augmentations that improve a model's bias to shape-based features rather than texture-based features results in increased OOD robustness for Convolutional Neural Networks trained on the ImageNet-1K dataset. This is usually stated as ``an increase in the model's shape bias results in an increase in its OOD robustness". Based on this hypothesis, some works in the literature aim to find augmentations with higher effects on model shape bias and use those for data augmentation. By evaluating 39 types of data augmentations on a widely used OOD dataset, we demonstrate the impact of each data augmentation on the model's robustness to OOD data and further show that the mentioned hypothesis is not true; an increase in shape bias does not necessarily result in higher OOD robustness. By analyzing the results, we also find some biases in the ImageNet-1K dataset that can easily be reduced using proper data augmentation. Our evaluation results further show that there is not necessarily a trade-off between in-domain accuracy and OOD robustness, and choosing the proper augmentations can help increase both in-domain accuracy and OOD robustness simultaneously.

PatchSVD: A Non-uniform SVD-based Image Compression Algorithm

Storing data is particularly a challenge when dealing with image data which often involves large file sizes due to the high resolution and complexity of images. Efficient image compression algorithms are crucial to better manage data storage costs. In this paper, we propose a novel region-based lossy image compression technique, called PatchSVD, based on the Singular Value Decomposition (SVD) algorithm. We show through experiments that PatchSVD outperforms SVD-based image compression with respect to three popular image compression metrics. Moreover, we compare PatchSVD compression artifacts with those of Joint Photographic Experts Group (JPEG) and SVD-based image compression and illustrate some cases where PatchSVD compression artifacts are preferable compared to JPEG and SVD artifacts.