Towards Open-Vocabulary Video Semantic Segmentation

Semantic segmentation in videos has been a focal point of recent research. However, existing models encounter challenges when faced with unfamiliar categories. To address this, we introduce the Open Vocabulary Video Semantic Segmentation (OV-VSS) task, designed to accurately segment every pixel across a wide range of open-vocabulary categories, including those that are novel or previously unexplored. To enhance OV-VSS performance, we propose a robust baseline, OV2VSS, which integrates a spatial-temporal fusion module, allowing the model to utilize temporal relationships across consecutive frames. Additionally, we incorporate a random frame enhancement module, broadening the model's understanding of semantic context throughout the entire video sequence. Our approach also includes video text encoding, which strengthens the model's capability to interpret textual information within the video context. Comprehensive evaluations on benchmark datasets such as VSPW and Cityscapes highlight OV-VSS's zero-shot generalization capabilities, especially in handling novel categories. The results validate OV2VSS's effectiveness, demonstrating improved performance in semantic segmentation tasks across diverse video datasets.

TimeSuite: Improving MLLMs for Long Video Understanding via Grounded Tuning

Multimodal Large Language Models (MLLMs) have demonstrated impressive performance in short video understanding. However, understanding long-form videos still remains challenging for MLLMs. This paper proposes TimeSuite, a collection of new designs to adapt the existing short-form video MLLMs for long video understanding, including a simple yet efficient framework to process long video sequence, a high-quality video dataset for grounded tuning of MLLMs, and a carefully-designed instruction tuning task to explicitly incorporate the grounding supervision in the traditional QA format. Specifically, based on VideoChat, we propose our long-video MLLM, coined as VideoChat-T, by implementing a token shuffling to compress long video tokens and introducing Temporal Adaptive Position Encoding (TAPE) to enhance the temporal awareness of visual representation. Meanwhile, we introduce the TimePro, a comprehensive grounding-centric instruction tuning dataset composed of 9 tasks and 349k high-quality grounded annotations. Notably, we design a new instruction tuning task type, called Temporal Grounded Caption, to peform detailed video descriptions with the corresponding time stamps prediction. This explicit temporal location prediction will guide MLLM to correctly attend on the visual content when generating description, and thus reduce the hallucination risk caused by the LLMs. Experimental results demonstrate that our TimeSuite provides a successful solution to enhance the long video understanding capability of short-form MLLM, achieving improvement of 5.6% and 6.8% on the benchmarks of Egoschema and VideoMME, respectively. In addition, VideoChat-T exhibits robust zero-shot temporal grounding capabilities, significantly outperforming the existing state-of-the-art MLLMs. After fine-tuning, it performs on par with the traditional supervised expert models.

VideoEval: Comprehensive Benchmark Suite for Low-Cost Evaluation of Video Foundation Model

With the growth of high-quality data and advancement in visual pre-training paradigms, Video Foundation Models (VFMs) have made significant progress recently, demonstrating their remarkable performance on traditional video understanding benchmarks. However, the existing benchmarks (e.g. Kinetics) and their evaluation protocols are often limited by relatively poor diversity, high evaluation costs, and saturated performance metrics. In this paper, we build a comprehensive benchmark suite to address these issues, namely VideoEval. Specifically, we establish the Video Task Adaption Benchmark (VidTAB) and the Video Embedding Benchmark (VidEB) from two perspectives: evaluating the task adaptability of VFMs under few-shot conditions and assessing their representation power by directly applying to downstream tasks. With VideoEval, we conduct a large-scale study on 20 popular open-source vision foundation models. Our study reveals some insightful findings on VFMs: 1) overall, current VFMs exhibit weak generalization across diverse tasks, 2) increasing video data, whether labeled or weakly-labeled video-text pairs, does not necessarily improve task performance, 3) the effectiveness of some pre-training paradigms may not be fully validated in previous benchmarks, and 4) combining different pre-training paradigms can help improve the generalization capabilities. We believe this study serves as an important complement to the current evaluation for VFMs and offers valuable insights for the future research.

Learning to (Learn at Test Time): RNNs with Expressive Hidden States

Self-attention performs well in long context but has quadratic complexity. Existing RNN layers have linear complexity, but their performance in long context is limited by the expressive power of their hidden state. We propose a new class of sequence modeling layers with linear complexity and an expressive hidden state. The key idea is to make the hidden state a machine learning model itself, and the update rule a step of self-supervised learning. Since the hidden state is updated by training even on test sequences, our layers are called Test-Time Training (TTT) layers. We consider two instantiations: TTT-Linear and TTT-MLP, whose hidden state is a linear model and a two-layer MLP respectively. We evaluate our instantiations at the scale of 125M to 1.3B parameters, comparing with a strong Transformer and Mamba, a modern RNN. Both TTT-Linear and TTT-MLP match or exceed the baselines. Similar to Transformer, they can keep reducing perplexity by conditioning on more tokens, while Mamba cannot after 16k context. With preliminary systems optimization, TTT-Linear is already faster than Transformer at 8k context and matches Mamba in wall-clock time. TTT-MLP still faces challenges in memory I/O, but shows larger potential in long context, pointing to a promising direction for future research.

Modelling the 5G Energy Consumption using Real-world Data: Energy Fingerprint is All You Need

The introduction of fifth-generation (5G) radio technology has revolutionized communications, bringing unprecedented automation, capacity, connectivity, and ultra-fast, reliable communications. However, this technological leap comes with a substantial increase in energy consumption, presenting a significant challenge. To improve the energy efficiency of 5G networks, it is imperative to develop sophisticated models that accurately reflect the influence of base station (BS) attributes and operational conditions on energy usage.Importantly, addressing the complexity and interdependencies of these diverse features is particularly challenging, both in terms of data processing and model architecture design. This paper proposes a novel 5G base stations energy consumption modelling method by learning from a real-world dataset used in the ITU 5G Base Station Energy Consumption Modelling Challenge in which our model ranked second. Unlike existing methods that omit the Base Station Identifier (BSID) information and thus fail to capture the unique energy fingerprint in different base stations, we incorporate the BSID into the input features and encoding it with an embedding layer for precise representation. Additionally, we introduce a novel masked training method alongside an attention mechanism to further boost the model's generalization capabilities and accuracy. After evaluation, our method demonstrates significant improvements over existing models, reducing Mean Absolute Percentage Error (MAPE) from 12.75% to 4.98%, leading to a performance gain of more than 60%.

InternVideo2: Scaling Video Foundation Models for Multimodal Video Understanding

We introduce InternVideo2, a new video foundation model (ViFM) that achieves the state-of-the-art performance in action recognition, video-text tasks, and video-centric dialogue. Our approach employs a progressive training paradigm that unifies the different self- or weakly-supervised learning frameworks of masked video token reconstruction, cross-modal contrastive learning, and next token prediction. Different training stages would guide our model to capture different levels of structure and semantic information through different pretext tasks. At the data level, we prioritize the spatiotemporal consistency by semantically segmenting videos and generating video-audio-speech captions. This improves the alignment between video and text. We scale both data and model size for our InternVideo2. Through extensive experiments, we validate our designs and demonstrate the state-of-the-art performance on over 60 video and audio tasks. Notably, our model outperforms others on various video-related captioning, dialogue, and long video understanding benchmarks, highlighting its ability to reason and comprehend long temporal contexts. Code and models are available at https://github.com/OpenGVLab/InternVideo2/.

VideoMamba: State Space Model for Efficient Video Understanding

Addressing the dual challenges of local redundancy and global dependencies in video understanding, this work innovatively adapts the Mamba to the video domain. The proposed VideoMamba overcomes the limitations of existing 3D convolution neural networks and video transformers. Its linear-complexity operator enables efficient long-term modeling, which is crucial for high-resolution long video understanding. Extensive evaluations reveal VideoMamba's four core abilities: (1) Scalability in the visual domain without extensive dataset pretraining, thanks to a novel self-distillation technique; (2) Sensitivity for recognizing short-term actions even with fine-grained motion differences; (3) Superiority in long-term video understanding, showcasing significant advancements over traditional feature-based models; and (4) Compatibility with other modalities, demonstrating robustness in multi-modal contexts. Through these distinct advantages, VideoMamba sets a new benchmark for video understanding, offering a scalable and efficient solution for comprehensive video understanding. All the code and models are available at https://github.com/OpenGVLab/VideoMamba.

Learning to (Learn at Test Time)

We reformulate the problem of supervised learning as learning to learn with two nested loops (i.e. learning problems). The inner loop learns on each individual instance with self-supervision before final prediction. The outer loop learns the self-supervised task used by the inner loop, such that its final prediction improves. Our inner loop turns out to be equivalent to linear attention when the inner-loop learner is only a linear model, and to self-attention when it is a kernel estimator. For practical comparison with linear or self-attention layers, we replace each of them in a transformer with an inner loop, so our outer loop is equivalent to training the architecture. When each inner-loop learner is a neural network, our approach vastly outperforms transformers with linear attention on ImageNet from 224 x 224 raw pixels in both accuracy and FLOPs, while (regular) transformers cannot run.

ZeroI2V: Zero-Cost Adaptation of Pre-trained Transformers from Image to Video

Adapting image models to video domain is becoming an efficient paradigm for solving video recognition tasks. Due to the huge number of parameters and effective transferability of image models, performing full fine-tuning is less efficient and even unnecessary. Thus, recent research is shifting its focus towards parameter-efficient image-to-video adaptation. However, these adaptation strategies inevitably introduce extra computational cost to deal with the domain gap and temporal modeling in videos. In this paper, our goal is to present a zero-cost adaptation paradigm (ZeroI2V) to transfer the image transformers to video recognition tasks (i.e., introduce zero extra cost to the adapted models during inference). To achieve this goal, we present two core designs. First, to capture the dynamics in videos and reduce the difficulty of achieving image-to-video adaptation, we exploit the flexibility of self-attention and introduce the spatial-temporal dual-headed attention (STDHA) that efficiently endow the image transformers with temporal modeling capability at zero extra parameters and computation. Second, to handle the domain gap between images and videos, we propose a linear adaption strategy which utilizes lightweight densely placed linear adapters to fully transfer the frozen image models to video recognition. Due to its customized linear design, all newly added adapters could be easily merged with the original modules through structural reparameterization after training, thus achieving zero extra cost during inference. Extensive experiments on four widely-used video recognition benchmarks show that our ZeroI2V can match or even outperform previous state-of-the-art methods while enjoying superior parameter and inference efficiency.

MIMOSA: Multi-constraint Molecule Sampling for Molecule Optimization

Molecule optimization is a fundamental task for accelerating drug discovery, with the goal of generating new valid molecules that maximize multiple drug properties while maintaining similarity to the input molecule. Existing generative models and reinforcement learning approaches made initial success, but still face difficulties in simultaneously optimizing multiple drug properties. To address such challenges, we propose the MultI-constraint MOlecule SAmpling (MIMOSA) approach, a sampling framework to use input molecule as an initial guess and sample molecules from the target distribution. MIMOSA first pretrains two property agnostic graph neural networks (GNNs) for molecule topology and substructure-type prediction, where a substructure can be either atom or single ring. For each iteration, MIMOSA uses the GNNs' prediction and employs three basic substructure operations (add, replace, delete) to generate new molecules and associated weights. The weights can encode multiple constraints including similarity and drug property constraints, upon which we select promising molecules for next iteration. MIMOSA enables flexible encoding of multiple property- and similarity-constraints and can efficiently generate new molecules that satisfy various property constraints and achieved up to 49.6% relative improvement over the best baseline in terms of success rate.