NeurIPS 2023 论文列表

Large language models (LLMs) have notably accelerated progress towards artificial general intelligence (AGI), with their impressive zero-shot capacity for user-tailored tasks, endowing them with immense potential across a range of applications. However, in the field of computer vision, despite the availability of numerous powerful vision foundation models (VFMs), they are still restricted to tasks in a pre-defined form, struggling to match the open-ended task capabilities of LLMs. In this work, we present an LLM-based framework for vision-centric tasks, termed VisionLLM. This framework provides a unified perspective for vision and language tasks by treating images as a foreign language and aligning vision-centric tasks with language tasks that can be flexibly defined and managed using language instructions. An LLM-based decoder can then make appropriate predictions based on these instructions for open-ended tasks. Extensive experiments show that the proposed VisionLLM can achieve different levels of task customization through language instructions, from fine-grained object-level to coarse-grained task-level customization, all with good results. It's noteworthy that, with a generalist LLM-based framework, our model can achieve over 60% mAP on COCO, on par with detection-specific models. We hope this model can set a new baseline for generalist vision and language models. The code shall be released.

The ubiquitous and demonstrably suboptimal choice of resizing images to a fixed resolution before processing them with computer vision models has not yet been successfully challenged. However, models such as the Vision Transformer (ViT) offer flexible sequence-based modeling, and hence varying input sequence lengths. We take advantage of this with NaViT (Native Resolution ViT) which uses sequence packing during training to process inputs of arbitrary resolutions and aspect ratios. Alongside flexible model usage, we demonstrate improved training efficiency for large-scale supervised and contrastive image-text pretraining. NaViT can be efficiently transferred to standard tasks such as image and video classification, object detection, and semantic segmentation and leads to improved results on robustness and fairness benchmarks. At inference time, the input resolution flexibility can be used to smoothly navigate the test-time cost-performance trade-off. We believe that NaViTmarks a departure from the standard, CNN-designed, input and modelling pipeline used by most computer vision models, and represents a promising direction for ViTs.

In this study, we investigate the task of data pre-selection, which aims to select instances for labeling from an unlabeled dataset through a single pass, thereby optimizing performance for undefined downstream tasks with a limited annotation budget. Previous approaches to data pre-selection relied solely on visual features extracted from foundation models, such as CLIP and BLIP-2, but largely ignored the powerfulness of text features. In this work, we argue that, with proper design, the joint feature space of both vision and text can yield a better representation for data pre-selection. To this end, we introduce UP-DP, a simple yet effective unsupervised prompt learning approach that adapts vision-language models, like BLIP-2, for data pre-selection. Specifically, with the BLIP-2 parameters frozen, we train text prompts to extract the joint features with improved representation, ensuring a diverse cluster structure that covers the entire dataset. We extensively compare our method with the state-of-the-art using seven benchmark datasets in different settings, achieving up to a performance gain of 20\%. Interestingly, the prompts learned from one dataset demonstrate significant generalizability and can be applied directly to enhance the feature extraction of BLIP-2 from other datasets. To the best of our knowledge, UP-DP is the first work to incorporate unsupervised prompt learning in a vision-language model for data pre-selection.

We propose a method to fuse frozen text-only large language models (LLMs) with pre-trained image encoder and decoder models, by mapping between their embedding spaces. Our model demonstrates a wide suite of multimodal capabilities: image retrieval, novel image generation, and multimodal dialogue. Ours is the first approach capable of conditioning on arbitrarily interleaved image and text inputs to generate coherent image (and text) outputs. To achieve strong performance on image generation, we propose an efficient mapping network to ground the LLM to an off-the-shelf text-to-image generation model. This mapping network translates hidden representations of text into the embedding space of the visual models, enabling us to leverage the strong text representations of the LLM for visual outputs. Our approach outperforms baseline generation models on tasks with longer and more complex language. In addition to novel image generation, our model is also capable of image retrieval from a prespecified dataset, and decides whether to retrieve or generate at inference time. This is done with a learnt decision module which conditions on the hidden representations of the LLM. Our model exhibits a wider range of capabilities compared to prior multimodal language models. It can process image-and-text inputs, and produce retrieved images, generated images, and generated text — outperforming non-LLM based generation models across several text-to-image tasks that measure context dependence.

Encoding models have been used to assess how the human brain represents concepts in language and vision. While language and vision rely on similar concept representations, current encoding models are typically trained and tested on brain responses to each modality in isolation. Recent advances in multimodal pretraining have produced transformers that can extract aligned representations of concepts in language and vision. In this work, we used representations from multimodal transformers to train encoding models that can transfer across fMRI responses to stories and movies. We found that encoding models trained on brain responses to one modality can successfully predict brain responses to the other modality, particularly in cortical regions that represent conceptual meaning. Further analysis of these encoding models revealed shared semantic dimensions that underlie concept representations in language and vision. Comparing encoding models trained using representations from multimodal and unimodal transformers, we found that multimodal transformers learn more aligned representations of concepts in language and vision. Our results demonstrate how multimodal transformers can provide insights into the brain’s capacity for multimodal processing.

Transformer-based language models have found many diverse applications requiring them to process sequences of increasing length. For these applications, the causal self-attention---which is the only component scaling quadratically w.r.t. the sequence length---becomes a central concern. While many works have proposed schemes to sparsify the attention patterns and reduce the computational overhead of self-attention, those are often limited by implementation concerns and end up imposing a simple and static structure over the attention matrix. Conversely, implementing more dynamic sparse attention often results in runtimes significantly slower than computing the full attention using the Flash implementation from Dao et al. (2022). We extend FlashAttention to accommodate a large class of attention sparsity patterns that, in particular, encompass key/query dropping and hashing-based attention. This leads to implementations with no computational complexity overhead and a multi-fold runtime speedup on top of FlashAttention. Even with relatively low degrees of sparsity, our method improves visibly upon FlashAttention as the sequence length increases. Without sacrificing perplexity, we increase the training speed of a transformer language model by $2.0\times$ and $3.3\times$ for sequences of respectively $8k$ and $16k$ tokens.

Pre-training on massive video datasets has become essential to achieve high action recognition performance on smaller downstream datasets. However, most large-scale video datasets contain images of people and hence are accompanied with issues related to privacy, ethics, and data protection, often preventing them from being publicly shared for reproducible research. Existing work has attempted to alleviate these problems by blurring faces, downsampling videos, or training on synthetic data. On the other hand, analysis on the {\em transferability} of privacy-preserving pre-trained models to downstream tasks has been limited. In this work, we study this problem by first asking the question: can we pre-train models for human action recognition with data that does not include real humans? To this end, we present, for the first time, a benchmark that leverages real-world videos with {\em humans removed} and synthetic data containing virtual humans to pre-train a model. We then evaluate the transferability of the representation learned on this data to a diverse set of downstream action recognition benchmarks. Furthermore, we propose a novel pre-training strategy, called Privacy-Preserving MAE-Align, to effectively combine synthetic data and human-removed real data. Our approach outperforms previous baselines by up to 5\% and closes the performance gap between human and no-human action recognition representations on downstream tasks, for both linear probing and fine-tuning. Our benchmark, code, and models are available at https://github.com/howardzh01/PPMA.

The contrastive vision-language pre-training, known as CLIP, demonstrates remarkable potential in perceiving open-world visual concepts, enabling effective zero-shot image recognition. Nevertheless, few-shot learning methods based on CLIP typically require offline fine-tuning of the parameters on few-shot samples, resulting in longer inference time and the risk of overfitting in certain domains. To tackle these challenges, we propose the Meta-Adapter, a lightweight residual-style adapter, to refine the CLIP features guided by the few-shot samples in an online manner. With a few training samples, our method can enable effective few-shot learning capabilities and generalize to unseen data or tasks without additional fine-tuning, achieving competitive performance and high efficiency. Without bells and whistles, our approach outperforms the state-of-the-art online few-shot learning method by an average of 3.6\% on eight image classification datasets with higher inference speed. Furthermore, our model is simple and flexible, serving as a plug-and-play module directly applicable to downstream tasks. Without further fine-tuning, Meta-Adapter obtains notable performance improvements in open-vocabulary object detection and segmentation tasks.

In this paper, we systematically evaluate the robustness of multi-exit language models against adversarial slowdown. To audit their robustness, we design a slowdown attack that generates natural adversarial text bypassing early-exit points. We use the resulting WAFFLE attack as a vehicle to conduct a comprehensive evaluation of three multi-exit mechanisms with the GLUE benchmark against adversarial slowdown. We then show our attack significantly reduces the computational savings provided by the three methods in both white-box and black-box settings. The more complex a mechanism is, the more vulnerable it is to adversarial slowdown. We also perform a linguistic analysis of the perturbed text inputs, identifying common perturbation patterns that our attack generates, and comparing them with standard adversarial text attacks. Moreover, we show that adversarial training is ineffective in defeating our slowdown attack, but input sanitization with a conversational model, e.g., ChatGPT, can remove perturbations effectively. This result suggests that future work is needed for developing efficient yet robust multi-exit models. Our code is available at: https://github.com/ztcoalson/WAFFLE

Deployed multimodal models can fail in ways that evaluators did not anticipate. In order to find these failures before deployment, we introduce MultiMon, a system that automatically identifies systematic failures---generalizable, natural-language descriptions that describe categories of individual failures. To uncover systematic failures, MultiMon scrapes for examples of erroneous agreement: inputs that produce the same output, but should not. It then prompts a language model to identify common categories and describe them in natural language. We use MultiMon to find 14 systematic failures (e.g."ignores quantifiers'') of the CLIP text-encoder, each comprising hundreds of distinct inputs (e.g."a shelf with a few/many books''). Because CLIP is the backbone for most state-of-the-art multimodal models, these inputs produce failures in Midjourney 5.1, DALL-E, VideoFusion, and others. MultiMon can also steer towards failures relevant to specific use cases, such as self-driving cars. We see MultiMon as a step towards evaluation that autonomously explores the long-tail of potential system failures.

The key to OOD detection has two aspects: generalized feature representation and precise category description. Recently, vision-language models such as CLIP provide significant advances in both two issues, but constructing precise category descriptions is still in its infancy due to the absence of unseen categories. This work introduces two hierarchical contexts, namely perceptual context and spurious context, to carefully describe the precise category boundary through automatic prompt tuning. Specifically, perceptual contexts perceive the inter-category difference (e.g., cats vs apples) for current classification tasks, while spurious contexts further identify spurious (similar but exactly not) OOD samples for every single category (e.g., cats vs panthers, apples vs peaches). The two contexts hierarchically construct the precise description for a certain category, which is, first roughly classifying a sample to the predicted category and then delicately identifying whether it is truly an ID sample or actually OOD. Moreover, the precise descriptions for those categories within the vision-language framework present a novel application: CATegory-EXtensible OOD detection (CATEX). One can efficiently extend the set of recognizable categories by simply merging the hierarchical contexts learned under different sub-task settings. And extensive experiments are conducted to demonstrate CATEX’s effectiveness, robustness, and category-extensibility. For instance, CATEX consistently surpasses the rivals by a large margin with several protocols on the challenging ImageNet-1K dataset. In addition, we offer new insights on how to efficiently scale up the prompt engineering in vision-language models to recognize thousands of object categories, as well as how to incorporate large language models (like GPT-3) to boost zero-shot applications.

One of the recent trends in vision problems is to use natural language captions to describe the objects of interest. This approach can overcome some limitations of traditional methods that rely on bounding boxes or category annotations. This paper introduces a novel paradigm for Multiple Object Tracking called Type-to-Track, which allows users to track objects in videos by typing natural language descriptions. We present a new dataset for that Grounded Multiple Object Tracking task, called GroOT, that contains videos with various types of objects and their corresponding textual captions describing their appearance and action in detail. Additionally, we introduce two new evaluation protocols and formulate evaluation metrics specifically for this task. We develop a new efficient method that models a transformer-based eMbed-ENcoDE-extRact framework (MENDER) using the third-order tensor decomposition. The experiments in five scenarios show that our MENDER approach outperforms another two-stage design in terms of accuracy and efficiency, up to 14.7\% accuracy and $4\times$ speed faster.

Transformers are central in modern natural language processing and computer vision applications. Despite recent works devoted to reducing the quadratic cost of such models with respect to sequence length, dealing with ultra long sequences (e.g., $>$16K tokens) remains challenging. Applications such as answering questions based on a book or summarizing a scientific article are inefficient or infeasible. Here, we propose to significantly improve the efficiency of Transformers for ultra long sequences, by compressing the sequence into a much smaller representation at each layer. Specifically, by exploiting the fact that in many tasks, only a small subset of special tokens, which we call VIP-tokens, are most relevant to the final prediction, we propose a VIP-token centric compression (VCC) scheme which selectively compresses the sequence based on their impact on approximating the representation of the VIP-tokens. Compared with competitive baselines, our algorithm is not only efficient (achieving more than $3\times$ compute efficiency gain compared to baselines on 4K and 16K lengths), but also offers competitive/better performance on a large number of tasks. Further, we show that our algorithm scales to 128K tokens (or more) while consistently offering accuracy improvement. Code is available at https://github.com/mlpen/VCC.

While Vision Transformers (ViTs) have undoubtedly made impressive strides in computer vision (CV), their intricate network structures necessitate substantial computation and memory resources. A decision-making process for CV tasks typically entails performing computations with low latency, which is a tricky problem for ViT models. Model quantization is a widely-used technique to optimize the hardware efficiency of deep neural networks. Full quantization under Sub-8-bit precision, in particular, is a promising solution to reduce inference latency significantly. Unfortunately, current commodity hardware, such as CPUs and GPUs, still struggles to efficiently execute these sub-8-bit quantized networks, as their SIMD instructions only support a granularity of 8 bits or wider. Also, there is a scarcity of literature that presents a full quantization paradigm for ViTs. In this paper, we propose an activation-aware fully sub-8-bit quantization-aware training (QAT) framework called PackQViT for efficient yet accurate ViT acceleration on mobile devices to facilitate real-time AI-powered decision-making. Specifically, in revisiting data activation within the ViT dataflow, two characteristics are relevant to quantization strategy and precision: the long-tailed distribution and systematic channel-wise outliers. In response, we employ either log2 quantization or clipping to address the long-tailed distribution and incorporate outlier-aware training for residual link quantization to regulate the various channel-wise outliers more consistently. Notably, due to the systematic fixed pattern, outlier-aware training approach can predict the channel indices and regularized scales of outliers in advance, thus avoiding the runtime data-adaptive selection during inference. Furthermore, we employ Int-$2^{n}$-Softmax, Int-LayerNorm, and Integer GELU to enable integer-only computation flow. Finally, we develop a SIMD-based 4-bit packed multiplier to achieve end-to-end ViT acceleration on mobile phones. Compared to prior studies on ViT quantization using 8-bit precision, PackQViT surpasses other works by an improved accuracy ranging from 0.4\% to 17.9\% for various widely used ViTs on ImageNet dataset; under 4-bit precision, PackQViT demonstrates 0.4%$\sim$2.8% higher accuracy. Compared to the baseline multiplier, our implementations on the Realme GT Android smartphone with Snapdragon 870 SoC CPU achieve 2.6x$\sim$3.7x speedup under 8-bit scenario and 3.8x$\sim$5.9x speedup under 4-bit which ensures practical real-time performance.

Pre-trained multi-modal vision-language models (VLMs) are becoming increasingly popular due to their exceptional performance on downstream vision applications, particularly in the few- and zero-shot settings. However, selecting the best-performing VLM for some downstream applications is non-trivial, as it is dataset and task-dependent. Meanwhile, the exhaustive evaluation of all available VLMs on a novel application is not only time and computationally demanding but also necessitates the collection of a labeled dataset for evaluation. As the number of open-source VLM variants increases, there is a need for an efficient model selection strategy that does not require access to a curated evaluation dataset. This paper proposes a novel task and benchmark for efficiently evaluating VLMs' zero-shot performance on downstream applications without access to the downstream task dataset. Specifically, we introduce a new task LOVM: **L**anguage-**O**nly **V**ision **M**odel Selection , where methods are expected to perform both model selection and performance prediction based solely on a text description of the desired downstream application. We then introduced an extensive LOVM benchmark consisting of ground-truth evaluations of 35 pre-trained VLMs and 23 datasets, where methods are expected to rank the pre-trained VLMs and predict their zero-shot performance.

In this paper, we provide the first efficient batched algorithm for contextual linear bandits with large action spaces. Unlike existing batched algorithms that rely on action elimination, which are not implementable for large action sets, our algorithm only uses a linear optimization oracle over the action set to design the policy. The proposed algorithm achieves a regret upper bound $\tilde{O}(\sqrt{T})$ with high probability, and uses $O(\log\log T)$ batches, matching the lower bound on the number of batches (Gao et al., 2019). When specialized to linear bandits, our algorithm can achieve a high probability gap-dependent regret bound of $\tilde{O}(1/\Delta_{\min})$ with the optimal $\log T$ number of batches, where $\Delta_{\min}$ is the minimum reward gap between a suboptimal arm and the optimal. Our result is achieved via a novel soft elimination approach, that entails $\text{``}$shaping$\text{"}$ the action sets at each batch so that we can efficiently identify (near) optimal actions.

Privacy-Preserving Action Recognition (PPAR) aims to transform raw videos into anonymous ones to prevent privacy leakage while maintaining action clues, which is an increasingly important problem in intelligent vision applications. Despite recent efforts in this task, it is still challenging to deal with novel privacy attributes and novel privacy attack models that are unavailable during the training phase. In this paper, from the perspective of meta-learning (learning to learn), we propose a novel Meta Privacy-Preserving Action Recognition (MPPAR) framework to improve both generalization abilities above (i.e., generalize to *novel privacy attributes* and *novel privacy attack models*) in a unified manner. Concretely, we simulate train/test task shifts by constructing disjoint support/query sets w.r.t. privacy attributes or attack models. Then, a virtual training and testing scheme is applied based on support/query sets to provide feedback to optimize the model's learning toward better generalization. Extensive experiments demonstrate the effectiveness and generalization of the proposed framework compared to state-of-the-arts.

We introduce EgoSchema, a very long-form video question-answering dataset, and benchmark to evaluate long video understanding capabilities of modern vision and language systems. Derived from Ego4D, EgoSchema consists of over 5000 human curated multiple choice question answer pairs, spanning over 250 hours of real video data, covering a very broad range of natural human activity and behavior. For each question, EgoSchema requires the correct answer to be selected between five given options based on a three-minute-long video clip. While some prior works have proposed video datasets with long clip lengths, we posit that merely the length of the video clip does not truly capture the temporal difficulty of the video task that is being considered. To remedy this, we introduce temporal certificate sets, a general notion for capturing the intrinsic temporal understanding length associated with a broad range of video understanding tasks & datasets. Based on this metric, we find EgoSchema to have intrinsic temporal lengths over 5.7x longer than the second closest dataset and 10x to 100x longer than any other video understanding dataset. Further, our evaluation of several current state-of-the-art video and language models shows them to be severely lacking in long-term video understanding capabilities. Even models with several billions of parameters achieve QA accuracy less than 33% (random is 20%) on the EgoSchema multi-choice question answering task, while humans achieve about 76% accuracy. We posit that EgoSchema, with its long intrinsic temporal structures and diverse complexity, would serve as a valuable evaluation probe for developing effective long-term video understanding systems in the future. Data and Zero-shot model evaluation code will all be open-sourced under the Ego4D license at http://egoschema.github.io.

Autoregressive transformers are spectacular models for short sequences but scale poorly to long sequences such as high-resolution images, podcasts, code, or books. We proposed Megabyte, a multi-scale decoder architecture that enables end-to-end differentiable modeling of sequences of over one million bytes. Megabyte segments sequences into patches and uses a local submodel within patches and a global model between patches. This enables sub-quadratic self-attention, much larger feedforward layers for the same compute, and improved parallelism during decoding---unlocking better performance at reduced cost for both training and generation. Extensive experiments show that Megabyte allows byte-level models to perform competitively with subword models on long context language modeling, achieve state-of-the-art density estimation on ImageNet, and model audio from raw files. Together, these results establish the viability of tokenization-free autoregressive sequence modeling at scale.