The rapid development of Artificial Intelligence (AI) has revolutionized numerous fields, with large language models (LLMs) and computer vision (CV) systems driving advancements in natural language understanding and visual processing, respectively. The convergence of these technologies has catalyzed the rise of multimodal AI, enabling richer, cross-modal understanding that spans text, vision, audio, and video modalities. Multimodal large language models (MLLMs), in particular, have emerged as a powerful framework, demonstrating impressive capabilities in tasks like image-text generation, visual question answering, and cross-modal retrieval. Despite these advancements, the complexity and scale of MLLMs introduce significant challenges in interpretability and explainability, essential for establishing transparency, trustworthiness, and reliability in high-stakes applications. This paper provides a comprehensive survey on the interpretability and explainability of MLLMs, proposing a novel framework that categorizes existing research across three perspectives: (I) Data, (II) Model, (III) Training \& Inference. We systematically analyze interpretability from token-level to embedding-level representations, assess approaches related to both architecture analysis and design, and explore training and inference strategies that enhance transparency. By comparing various methodologies, we identify their strengths and limitations and propose future research directions to address unresolved challenges in multimodal explainability. This survey offers a foundational resource for advancing interpretability and transparency in MLLMs, guiding researchers and practitioners toward developing more accountable and robust multimodal AI systems.

The advent and proliferation of large multi-modal models (LMMs) have introduced new paradigms to computer vision, transforming various tasks into a unified visual question answering framework. Video Quality Assessment (VQA), a classic field in low-level visual perception, focused initially on quantitative video quality scoring. However, driven by advances in LMMs, it is now progressing toward more holistic visual quality understanding tasks. Recent studies in the image domain have demonstrated that Visual Question Answering (VQA) can markedly enhance low-level visual quality evaluation. Nevertheless, related work has not been explored in the video domain, leaving substantial room for improvement. To address this gap, we introduce the VQA2 Instruction Dataset - the first visual question answering instruction dataset that focuses on video quality assessment. This dataset consists of 3 subsets and covers various video types, containing 157,755 instruction question-answer pairs. Then, leveraging this foundation, we present the VQA2 series models. The VQA2 series models interleave visual and motion tokens to enhance the perception of spatial-temporal quality details in videos. We conduct extensive experiments on video quality scoring and understanding tasks, and results demonstrate that the VQA2series models achieve excellent performance in both tasks. Notably, our final model, the VQA2-Assistant, exceeds the renowned GPT-4o in visual quality understanding tasks while maintaining strong competitiveness in quality scoring tasks. Our work provides a foundation and feasible approach for integrating low-level video quality assessment and understanding with LMMs.

Recently, the quality of artworks generated using Artificial Intelligence (AI) has increased significantly, resulting in growing difficulties in detecting synthetic artworks. However, limited studies have been conducted on identifying the authenticity of synthetic artworks and their source. This paper introduces AI-ArtBench, a dataset featuring 185,015 artistic images across 10 art styles. It includes 125,015 AI-generated images and 60,000 pieces of human-created artwork. This paper also outlines a method to accurately detect AI-generated images and trace them to their source model. This work proposes a novel Convolutional Neural Network model based on the ConvNeXt model called AttentionConvNeXt. AttentionConvNeXt was implemented and trained to differentiate between the source of the artwork and its style with an F1-Score of 0.869. The accuracy of attribution to the generative model reaches 0.999. To combine the scientific contributions arising from this study, a web-based application named ArtBrain was developed to enable both technical and non-technical users to interact with the model. Finally, this study presents the results of an Artistic Turing Test conducted with 50 participants. The findings reveal that humans could identify AI-generated images with an accuracy of approximately 58%, while the model itself achieved a significantly higher accuracy of around 99%.

We introduce Presto, a novel video diffusion model designed to generate 15-second videos with long-range coherence and rich content. Extending video generation methods to maintain scenario diversity over long durations presents significant challenges. To address this, we propose a Segmented Cross-Attention (SCA) strategy, which splits hidden states into segments along the temporal dimension, allowing each segment to cross-attend to a corresponding sub-caption. SCA requires no additional parameters, enabling seamless incorporation into current DiT-based architectures. To facilitate high-quality long video generation, we build the LongTake-HD dataset, consisting of 261k content-rich videos with scenario coherence, annotated with an overall video caption and five progressive sub-captions. Experiments show that our Presto achieves 78.5% on the VBench Semantic Score and 100% on the Dynamic Degree, outperforming existing state-of-the-art video generation methods. This demonstrates that our proposed Presto significantly enhances content richness, maintains long-range coherence, and captures intricate textual details. More details are displayed on our project page: https://presto-video.github.io/.

Emotion recognition is a critical component of affective computing. Training accurate machine learning models for emotion recognition typically requires a large amount of labeled data. Due to the subtleness and complexity of emotions, multiple evaluators are usually needed for each affective sample to obtain its ground-truth label, which is expensive. To save the labeling cost, this paper proposes an inconsistency-based active learning approach for cross-task transfer between emotion classification and estimation. Affective norms are utilized as prior knowledge to connect the label spaces of categorical and dimensional emotions. Then, the prediction inconsistency on the two tasks for the unlabeled samples is used to guide sample selection in active learning for the target task. Experiments on within-corpus and cross-corpus transfers demonstrated that cross-task inconsistency could be a very valuable metric in active learning. To our knowledge, this is the first work that utilizes prior knowledge on affective norms and data in a different task to facilitate active learning for a new task, even the two tasks are from different datasets.

Current semantic segmentation models typically require a substantial amount of manually annotated data, a process that is both time-consuming and resource-intensive. Alternatively, leveraging advanced text-to-image models such as Midjourney and Stable Diffusion has emerged as an efficient strategy, enabling the automatic generation of synthetic data in place of manual annotations. However, previous methods have been limited to generating single-instance images, as the generation of multiple instances with Stable Diffusion has proven unstable. To address this limitation and expand the scope and diversity of synthetic datasets, we propose a framework \textbf{Free-Mask} that combines a Diffusion Model for segmentation with advanced image editing capabilities, allowing for the integration of multiple objects into images via text-to-image models. Our method facilitates the creation of highly realistic datasets that closely emulate open-world environments while generating accurate segmentation masks. It reduces the labor associated with manual annotation and also ensures precise mask generation. Experimental results demonstrate that synthetic data generated by \textbf{Free-Mask} enables segmentation models to outperform those trained on real data, especially in zero-shot settings. Notably, \textbf{Free-Mask} achieves new state-of-the-art results on previously unseen classes in the VOC 2012 benchmark.

Lifelike visualizations in design, cinematography, and gaming rely on precise physics simulations, typically requiring extensive computational resources and detailed physical input. This paper presents a method that can infer a system's physical properties from a short video, eliminating the need for explicit parameter input, provided it is close to the training condition. The learned representation is then used within a Graph Network-based Simulator to emulate the trajectories of physical systems. We demonstrate that the video-derived encodings effectively capture the physical properties of the system and showcase a linear dependence between some of the encodings and the system's motion.

Most tech costs a fair chunk of change (the PS5 Pro and iPad Pro jump to mind), but some smaller electronics are (almost) as exciting. Now that Cyber Monday is here, we've found dozens of deals on inexpensive electronics and accessories that cost less than 50. Unlike other lists you may find today, this one is made up of items we've tested, reviewed or otherwise recommend in our buying and gift guides -- so we think they're actually worth your money. Amazon Echo Pop (2023) for 18 ( 22 off): Amazon's smallest Echo will fit in any room in your home, so Alexa can add things to your shopping list, set a timer, or answer questions (like "What's a bomb cyclone?" or "Who is Penelope Cruz married to?") from anywhere. Anker Nano Charger 30W USB-C for 13 ( 7 off): This compact 30-watt wall charger is smaller than others of its wattage and can speedily juice up an iPhone or Android handset. Anker is one of Engadget's most recommended accessory brands and this is the model we picked for our fast charger guide. Get the same deal at Anker with an auto-applied code. Anker Nano power bank with built-in USB-C connector for 16 ( 4 off): It's the size of an old-timey lipstick case but packs enough juice (and its own USB-C plug) to get a dying smartphone back in service with at least a half charge.

The expensive tech gets all the attention -- thousand-dollar phones and 500 tablets. But the supporting players, the cables and batteries and chargers that make those devices work properly, are just as important. Right now for Black Friday, many of those smaller gadgets are on sale for less than 50 even after the day has passed. And there are even some standalone devices like earbuds and smart speakers that fall well below the threshold. We've tested scads of these smaller, less expensive tech for Engadget buying guides, including the best power banks, iPad accessories, smart plugs and microSD cards. Here, we've gathered up all the Black Friday tech deals under 50 that you can still get on gadgets we recommend. Amazon Echo Pop (2023) for 18 ( 22 off): Amazon's smallest Echo will fit in any room in your home, so Alexa can add things to your shopping list, set a timer, or answer questions (like "What's a bomb cyclone?" or "Who is Penelope Cruz married to?") from anywhere.

Large language models (LLMs) are often sycophantic, prioritizing agreement with their users over accurate or objective statements. This problematic behavior becomes more pronounced during reinforcement learning from human feedback (RLHF), an LLM fine-tuning stage intended to align model outputs with human values. Instead of increasing accuracy and reliability, the reward model learned from RLHF often rewards sycophancy. We develop a linear probing method to identify and penalize markers of sycophancy within the reward model, producing rewards that discourage sycophantic behavior. Our experiments show that constructing and optimizing against this surrogate reward function reduces sycophantic behavior in multiple open-source LLMs. Our results suggest a generalizable methodology for reducing unwanted LLM behaviors that are not sufficiently disincentivized by RLHF fine-tuning.