Large Language Models (LLMs) can become outdated over time as they may lack updated world knowledge, leading to factual knowledge errors and gaps. Knowledge Editing (KE) aims to overcome this challenge using weight updates that do not require expensive retraining. We propose treating KE as an LLM alignment problem. Toward this goal, we introduce Knowledge Direct Preference Optimization (KDPO), a variation of the Direct Preference Optimization (DPO) that is more effective for knowledge modifications. Our method is based on an online approach that continually updates the knowledge stored in the model. We use the current knowledge as a negative sample and the new knowledge we want to introduce as a positive sample in a process called DPO. We also use teacher-forcing for negative sample generation and optimize using the positive sample, which helps maintain localized changes. We tested our KE method on various datasets and models, comparing it to several cutting-edge methods, with 100 and 500 sequential edits. Additionally, we conducted an ablation study comparing our method to the standard DPO approach. Our experimental results show that our modified DPO method allows for more refined KE, achieving similar or better performance compared to previous methods.

Large-scale text-to-image models that can generate high-quality and diverse images based on textual prompts have shown remarkable success. These models aim ultimately to create complex scenes, and addressing the challenge of multi-subject generation is a critical step towards this goal. However, the existing state-of-the-art diffusion models face difficulty when generating images that involve multiple subjects. When presented with a prompt containing more than one subject, these models may omit some subjects or merge them together. To address this challenge, we propose a novel approach based on a guiding principle. We allow the diffusion model to initially propose a layout, and then we rearrange the layout grid. This is achieved by enforcing cross-attention maps (XAMs) to adhere to proposed masks and by migrating pixels from latent maps to new locations determined by us. We introduce new loss terms aimed at reducing XAM entropy for clearer spatial definition of subjects, reduce the overlap between XAMs, and ensure that XAMs align with their respective masks. We contrast our approach with several alternative methods and show that it more faithfully captures the desired concepts across a variety of text prompts.

Density estimation based anomaly detection schemes typically model anomalies as examples that reside in low-density regions. We propose a modified density estimation problem and demonstrate its effectiveness for anomaly detection. Specifically, we assume the density function of normal samples is uniform in some compact domain. This assumption implies the density function is more stable (with lower variance) around normal samples than anomalies. We first corroborate this assumption empirically using a wide range of real-world data. Then, we design a variance stabilized density estimation problem for maximizing the likelihood of the observed samples while minimizing the variance of the density around normal samples. We introduce an ensemble of autoregressive models to learn the variance stabilized distribution. Finally, we perform an extensive benchmark with 52 datasets demonstrating that our method leads to state-of-the-art results while alleviating the need for data-specific hyperparameter tuning.

The stochastic gradient noise (SGN) is a significant factor in the success of stochastic gradient descent (SGD). Following the central limit theorem, SGN was initially modeled as Gaussian, and lately, it has been suggested that stochastic gradient noise is better characterized using $S\alpha S$ L\'evy distribution. This claim was allegedly refuted and rebounded to the previously suggested Gaussian noise model. This paper presents solid, detailed empirical evidence that SGN is heavy-tailed and better depicted by the $S\alpha S$ distribution. Furthermore, we argue that different parameters in a deep neural network (DNN) hold distinct SGN characteristics throughout training. To more accurately approximate the dynamics of SGD near a local minimum, we construct a novel framework in $\mathbb{R}^N$, based on L\'evy-driven stochastic differential equation (SDE), where one-dimensional L\'evy processes model each parameter in the DNN. Next, we show that SGN jump intensity (frequency and amplitude) depends on the learning rate decay mechanism (LRdecay); furthermore, we demonstrate empirically that the LRdecay effect may stem from the reduction of the SGN and not the decrease in the step size. Based on our analysis, we examine the mean escape time, trapping probability, and more properties of DNNs near local minima. Finally, we prove that the training process will likely exit from the basin in the direction of parameters with heavier tail SGN. We will share our code for reproducibility.

Accurately clustering high-dimensional measurements is vital for adequately analyzing scientific data. Deep learning machinery has remarkably improved clustering capabilities in recent years due to its ability to extract meaningful representations. In this work, we are given unlabeled samples from multiple source domains, and we aim to learn a shared classifier that assigns the examples to various clusters. Evaluation is done by using the classifier for predicting cluster assignments in a previously unseen domain. This setting generalizes the problem of unsupervised domain generalization to the case in which no supervised learning samples are given (completely unsupervised). Towards this goal, we present an end-to-end model and evaluate its capabilities on several multi-domain image datasets. Specifically, we demonstrate that our model is more accurate than schemes that require fine-tuning using samples from the target domain or some level of supervision.

Feature whitening is a known technique for speeding up training of DNN. Under certain assumptions, whitening the activations reduces the Fisher information matrix to a simple identity matrix, in which case stochastic gradient descent is equivalent to the faster natural gradient descent. Due to the additional complexity resulting from transforming the layer inputs and their corresponding gradients in the forward and backward propagation, and from repeatedly computing the Eigenvalue decomposition (EVD), this method is not commonly used to date. In this work, we address the complexity drawbacks of feature whitening. Our contribution is twofold. First, we derive an equivalent method, which replaces the sample transformations by a transformation to the weight gradients, applied to every batch of B samples. The complexity is reduced by a factor of S=(2B), where S denotes the feature dimension of the layer output. As the batch size increases with distributed training, the benefit of using the proposed method becomes more compelling. Second, motivated by the theoretical relation between the condition number of the sample covariance matrix and the convergence speed, we derive an alternative sub-optimal algorithm which recursively reduces the condition number of the latter matrix. Compared to EVD, complexity is reduced by a factor of the input feature dimension M. We exemplify the proposed algorithms with ResNet-based networks for image classification demonstrated on the CIFAR and Imagenet datasets. Parallelizing the proposed algorithms is straightforward and we implement a distributed version thereof. Improved convergence, in terms of speed and attained accuracy, can be observed in our experiments.

Adaptive and Iteratively Improving Recurrent Lateral ConnectionsBarak Battash Lior Wolf Tel Aviv University Facebook AI Research & Tel Aviv University Abstract The current leading computer vision models are typically feed forward neural models, in which the output of one computational block is passed to the next one sequentially. This is in sharp contrast to the organization of the primate visual cortex, in which feedback and lateral connections are abundant. In this work, we propose a computational model for the role of lateral connections in a given block, in which the weights of the block vary dynamically as a function of its activations, and the input from the upstream blocks is iteratively reintroduced. We demonstrate how this novel architectural modification can lead to sizable gains in performance, when applied to visual action recognition without pretraining and that it outperforms the literature architectures with recurrent feedback processing on ImageNet. 1 Introduction Rapid exposure experiments in primates teach us that image recognition occurs as early as the first 100 msec of visual perception, a time budget that suffices only for feed-forward inference, due to the relatively slow nature of biological neurons (Perrett and Oram, 1993; Thorpe et al., 1996). However, anatomical studies have shown that feedback connections are prevalent in the cortex (Douglas and Martin, 2004; Felleman and Essen, 1991). As one striking example, the feedforward input from LGN to V1 in cats constitutes only five percent of the total input to V1, the rest being lateral and feedback connections (Binzegger et al., 2004). In fact, lateral connections, which are projections from a layer to itself, are even more prevalent than feedback connections that project from downstream layers upstream. One possible conjecture would be that feedback (including lateral) connections play roles that are replaced by other mechanisms in the current deep learning literature. For example, they could play a role in training the biological neural network, or they can form attention mechanisms, which are captured by attention (Sermanet et al., 2014) and self-attention (Parikh et al., 2016) blocks in modern neural networks. Similarly, one can claim that such connections are required due to the limitations of the biological computational, but may not be necessary in artificial neural networks, which can be extremely deep (Liao and Poggio, 2016).