MEDSYN: Benchmarking Multi-EviDence SYNthesis in Complex Clinical Cases for Multimodal Large Language Models

Multimodal large language models (MLLMs) have shown great potential in medical applications, yet existing benchmarks inadequately capture real-world clinical complexity. We introduce MEDSYN, a multilingual, multimodal benchmark of highly complex clinical cases with up to 7 distinct visual clinical evidence (CE) types per case. Mirroring clinical workflow, we evaluate 18 MLLMs on differential diagnosis (DDx) generation and final diagnosis (FDx) selection. While top models often match or even outperform human experts on DDx generation, all MLLMs exhibit a much larger DDx--FDx performance gap compared to expert clinicians, indicating a failure mode in synthesis of heterogeneous CE types. Ablations attribute this failure to (i) overreliance on less discriminative textual CE ($\it{e.g.}$, medical history) and (ii) a cross-modal CE utilization gap. We introduce Evidence Sensitivity to quantify the latter and show that a smaller gap correlates with higher diagnostic accuracy. Finally, we demonstrate how it can be used to guide interventions to improve model performance. We will open-source our benchmark and code.

Mask What Matters: Mitigating Object Hallucinations in Multimodal Large Language Models with Object-Aligned Visual Contrastive Decoding

We study object hallucination in Multimodal Large Language Models (MLLMs) and improve visual contrastive decoding (VCD) by constructing an object-aligned auxiliary view. We leverage object-centric attention in self-supervised Vision Transformers. In particular, we remove the most salient visual evidence to construct an auxiliary view that disrupts unsupported tokens and produces a stronger contrast signal. Our method is prompt-agnostic, model-agnostic, and can be seamlessly plugged into the existing VCD pipeline with little computation overhead, i.e., a single cacheable forward pass. Empirically, our method demonstrates consistent gains on two popular object hallucination benchmarks across two MLLMs.

Towards Robust Protective Perturbation against DeepFake Face Swapping

DeepFake face swapping enables highly realistic identity forgeries, posing serious privacy and security risks. A common defence embeds invisible perturbations into images, but these are fragile and often destroyed by basic transformations such as compression or resizing. In this paper, we first conduct a systematic analysis of 30 transformations across six categories and show that protection robustness is highly sensitive to the choice of training transformations, making the standard Expectation over Transformation (EOT) with uniform sampling fundamentally suboptimal. Motivated by this, we propose Expectation Over Learned distribution of Transformation (EOLT), the framework to treat transformation distribution as a learnable component rather than a fixed design choice. Specifically, EOLT employs a policy network that learns to automatically prioritize critical transformations and adaptively generate instance-specific perturbations via reinforcement learning, enabling explicit modeling of defensive bottlenecks while maintaining broad transferability. Extensive experiments demonstrate that our method achieves substantial improvements over state-of-the-art approaches, with 26% higher average robustness and up to 30% gains on challenging transformation categories.

SynapseRoute: An Auto-Route Switching Framework on Dual-State Large Language Model

With the widespread adoption of large language models (LLMs) in practical applications, selecting an appropriate model requires balancing not only performance but also operational cost. The emergence of reasoning-capable models has further widened the cost gap between "thinking" (high reasoning) and "non-thinking" (fast, low-cost) modes. In this work, we reveal that approximately 58% of medical questions can be accurately answered by the non-thinking mode alone, without requiring the high-cost reasoning process. This highlights a clear dichotomy in problem complexity and suggests that dynamically routing queries to the appropriate mode based on complexity could optimize accuracy, cost-efficiency, and overall user experience. Based on this, we further propose SynapseRoute, a machine learning-based dynamic routing framework that intelligently assigns input queries to either thinking or non-thinking modes. Experimental results on several medical datasets demonstrate that SynapseRoute not only improves overall accuracy (0.8390 vs. 0.8272) compared to the thinking mode alone but also reduces inference time by 36.8% and token consumption by 39.66%. Importantly, qualitative analysis indicates that over-reasoning on simpler queries can lead to unnecessary delays and even decreased accuracy, a pitfall avoided by our adaptive routing. Finally, this work further introduces the Accuracy-Inference-Token (AIT) index to comprehensively evaluate the trade-offs among accuracy, latency, and token cost.

Artificial Intelligence for Biomedical Video Generation

As a prominent subfield of Artificial Intelligence Generated Content (AIGC), video generation has achieved notable advancements in recent years. The introduction of Sora-alike models represents a pivotal breakthrough in video generation technologies, significantly enhancing the quality of synthesized videos. Particularly in the realm of biomedicine, video generation technology has shown immense potential such as medical concept explanation, disease simulation, and biomedical data augmentation. In this article, we thoroughly examine the latest developments in video generation models and explore their applications, challenges, and future opportunities in the biomedical sector. We have conducted an extensive review and compiled a comprehensive list of datasets from various sources to facilitate the development and evaluation of video generative models in biomedicine. Given the rapid progress in this field, we have also created a github repository to regularly update the advances of biomedical video generation at: https://github.com/Lee728243228/Biomedical-Video-Generation

MEDCO: Medical Education Copilots Based on A Multi-Agent Framework

Large language models (LLMs) have had a significant impact on diverse research domains, including medicine and healthcare. However, the potential of LLMs as copilots in medical education remains underexplored. Current AI-assisted educational tools are limited by their solitary learning approach and inability to simulate the multi-disciplinary and interactive nature of actual medical training. To address these limitations, we propose MEDCO (Medical EDucation COpilots), a novel multi-agent-based copilot system specially developed to emulate real-world medical training environments. MEDCO incorporates three primary agents: an agentic patient, an expert doctor, and a radiologist, facilitating a multi-modal and interactive learning environment. Our framework emphasizes the learning of proficient question-asking skills, multi-disciplinary collaboration, and peer discussions between students. Our experiments show that simulated virtual students who underwent training with MEDCO not only achieved substantial performance enhancements comparable to those of advanced models, but also demonstrated human-like learning behaviors and improvements, coupled with an increase in the number of learning samples. This work contributes to medical education by introducing a copilot that implements an interactive and collaborative learning approach. It also provides valuable insights into the effectiveness of AI-integrated training paradigms.

One Prompt Word is Enough to Boost Adversarial Robustness for Pre-trained Vision-Language Models

Large pre-trained Vision-Language Models (VLMs) like CLIP, despite having remarkable generalization ability, are highly vulnerable to adversarial examples. This work studies the adversarial robustness of VLMs from the novel perspective of the text prompt instead of the extensively studied model weights (frozen in this work). We first show that the effectiveness of both adversarial attack and defense are sensitive to the used text prompt. Inspired by this, we propose a method to improve resilience to adversarial attacks by learning a robust text prompt for VLMs. The proposed method, named Adversarial Prompt Tuning (APT), is effective while being both computationally and data efficient. Extensive experiments are conducted across 15 datasets and 4 data sparsity schemes (from 1-shot to full training data settings) to show APT's superiority over hand-engineered prompts and other state-of-the-art adaption methods. APT demonstrated excellent abilities in terms of the in-distribution performance and the generalization under input distribution shift and across datasets. Surprisingly, by simply adding one learned word to the prompts, APT can significantly boost the accuracy and robustness (epsilon=4/255) over the hand-engineered prompts by +13% and +8.5% on average respectively. The improvement further increases, in our most effective setting, to +26.4% for accuracy and +16.7% for robustness. Code is available at https://github.com/TreeLLi/APT.

A Survey of Reasoning with Foundation Models

Reasoning, a crucial ability for complex problem-solving, plays a pivotal role in various real-world settings such as negotiation, medical diagnosis, and criminal investigation. It serves as a fundamental methodology in the field of Artificial General Intelligence (AGI). With the ongoing development of foundation models, there is a growing interest in exploring their abilities in reasoning tasks. In this paper, we introduce seminal foundation models proposed or adaptable for reasoning, highlighting the latest advancements in various reasoning tasks, methods, and benchmarks. We then delve into the potential future directions behind the emergence of reasoning abilities within foundation models. We also discuss the relevance of multimodal learning, autonomous agents, and super alignment in the context of reasoning. By discussing these future research directions, we hope to inspire researchers in their exploration of this field, stimulate further advancements in reasoning with foundation models, and contribute to the development of AGI.

Dietary Assessment with Multimodal ChatGPT: A Systematic Analysis

Conventional approaches to dietary assessment are primarily grounded in self-reporting methods or structured interviews conducted under the supervision of dietitians. These methods, however, are often subjective, potentially inaccurate, and time-intensive. Although artificial intelligence (AI)-based solutions have been devised to automate the dietary assessment process, these prior AI methodologies encounter challenges in their ability to generalize across a diverse range of food types, dietary behaviors, and cultural contexts. This results in AI applications in the dietary field that possess a narrow specialization and limited accuracy. Recently, the emergence of multimodal foundation models such as GPT-4V powering the latest ChatGPT has exhibited transformative potential across a wide range of tasks (e.g., Scene understanding and image captioning) in numerous research domains. These models have demonstrated remarkable generalist intelligence and accuracy, capable of processing various data modalities. In this study, we explore the application of multimodal ChatGPT within the realm of dietary assessment. Our findings reveal that GPT-4V excels in food detection under challenging conditions with accuracy up to 87.5% without any fine-tuning or adaptation using food-specific datasets. By guiding the model with specific language prompts (e.g., African cuisine), it shifts from recognizing common staples like rice and bread to accurately identifying regional dishes like banku and ugali. Another GPT-4V's standout feature is its contextual awareness. GPT-4V can leverage surrounding objects as scale references to deduce the portion sizes of food items, further enhancing its accuracy in translating food weight into nutritional content. This alignment with the USDA National Nutrient Database underscores GPT-4V's potential to advance nutritional science and dietary assessment techniques.

Aria-NeRF: Multimodal Egocentric View Synthesis

We seek to accelerate research in developing rich, multimodal scene models trained from egocentric data, based on differentiable volumetric ray-tracing inspired by Neural Radiance Fields (NeRFs). The construction of a NeRF-like model from an egocentric image sequence plays a pivotal role in understanding human behavior and holds diverse applications within the realms of VR/AR. Such egocentric NeRF-like models may be used as realistic simulations, contributing significantly to the advancement of intelligent agents capable of executing tasks in the real-world. The future of egocentric view synthesis may lead to novel environment representations going beyond today's NeRFs by augmenting visual data with multimodal sensors such as IMU for egomotion tracking, audio sensors to capture surface texture and human language context, and eye-gaze trackers to infer human attention patterns in the scene. To support and facilitate the development and evaluation of egocentric multimodal scene modeling, we present a comprehensive multimodal egocentric video dataset. This dataset offers a comprehensive collection of sensory data, featuring RGB images, eye-tracking camera footage, audio recordings from a microphone, atmospheric pressure readings from a barometer, positional coordinates from GPS, connectivity details from Wi-Fi and Bluetooth, and information from dual-frequency IMU datasets (1kHz and 800Hz) paired with a magnetometer. The dataset was collected with the Meta Aria Glasses wearable device platform. The diverse data modalities and the real-world context captured within this dataset serve as a robust foundation for furthering our understanding of human behavior and enabling more immersive and intelligent experiences in the realms of VR, AR, and robotics.