#MultimodalAI

Multimodal AI is changing how systems understand the world not just text. which multimodal AI topic is hardest to master? #MultimodalAI #GenerativeAI #AI #MachineLearning #LLM #Examadda

🤖 Multimodal AI in action! 🎭📊🎧 From Audio & Visual QA to 3D Synthesis, Video QA, and Math Reasoning, AI is mastering diverse tasks across modalities. 🚀 How far can AI go in seeing, hearing & thinking? 🔥 #AI #MachineLearning #MultimodalAI #DeepLearning

We introduce PhenoTrait, a multimodal plant disease diagnosis model that outperforms GPT-4o by aligning images with expert text. Code & data public. #PlantDisease #MultimodalAI #OpenScience Details:

Heading to #ICLR2026 in Rio de Janeiro this week! 🇧🇷🏖️ Would love to connect with folks in Multimodal AI, AI Agents, Graph Learning, or Computer-Use Agents — or just grab a coffee/caipirinha! ☕️🍹 Presenting 4 papers 👇 (2 main + 2 workshops)

4/ SCOPE 🔍🎨 Selective Cross-modal Orchestration of Visual Perception Experts — smarter, selective orchestration of visual experts in multimodal models. 📍 Room 204 C · Workshop on Multimodal Intelligence 🗓 Sun Apr 26 · 2:30 – 4:00 PM BRT 🔗

Huge thanks to all my brilliant co-authors and mentors 🙌🚀 If you're at ICLR and want to chat AI Agents, Multimodal Learning, Graph Learning, CUAs, or anything else — please stop by our posters or DM me! See you in Rio 🇧🇷🌊⛱️ #AI #Agents #MultimodalAI #GraphLearning #LLM #CUA

The future of AI is multimodal and Baidu’s ERNIE 5.0 just threw down the gauntlet against GPT-5. This isn’t just an upgrade; it’s rewriting how AI understands the world. Here’s why it matters. 👇 #MultimodalAI #AI 1/9 🧵

Our portco @Vectara adapted Mockingbird V3 MM for #Semiconductors, raising factual consistency on engineering images from 0.29 to 0.45. Specialized #MultimodalAI helps engineers read diagrams and get more reliable answers. #AIResearch #EnterpriseAI 🔗:

🚨Qwen Work just added AI multimodal generation. You can now create marketing posters, product videos, and brand music directly inside the app. One workspace, from visual content to video and audio creation. #QwenWork #MultimodalAI

By 2030, 80% of enterprise software will be multimodal — reshaping user experiences and automation. Future-proof your agentic AI investments and prepare for multiagent systems: #GartnerIT #GenAI #MultimodalAI #Innovation

— Multimodal Unified AI 🧠 One Intelligence. Every Modality. 3L. si A powerful name for the next generation of multimodal intelligence. #AI #MultimodalAI #ArtificialIntelligence #Domains #DomaineforSale

AI can now work across text, images, audio, and video. 🤖 Multimodal AI models connect different types of information to make AI more useful for real-world tasks. 👇 #MultimodalAI #ArtificialIntelligence #AI #TechTrends #TechnologyMoment

#multimodalai project ideas to boost your resume: 1. Image Captioning and Recommendation System 2. Automatic Speech Recognition (ASR) 3. Visual Question Answering (VQA) Find solved & explained examples here: #ArtificialIntelligence #MachineLearning

Meet the reComputer @NVIDIARobotics #AGXOrin series: a supercomputing edge platform that turns #multimodalAI into action. Up to 275 TOPS AI perfromance, feeds from 8x #GMSL cameras for full-surround vision, togehter with Industrial I/O (#CAN FD, #10G Ethernet, wide voltage) for reliable deployment. Perfect to runs real time #3D mapping, multimodal #AIAgents, and dexterous manipulation.

Multimodal Edge AI = 👁️ See 👂 Hear ✋ Feel 🧠 Understand When devices sense like humans, the world changes. #MultimodalAI #EdgeAI

How can AI responsibly deepen our understanding of human behaviors?🧠✨ Join us at #CSCW2024 for our Special Interest Group where we welcome interdisciplinary voices! 📅 Nov 11, 11 AM @YunHuang_HCI, @QiaoJin5, @Xi_Lu_HCI, @praeclarumjj, @humphrey_shi #MultimodalAI #EthicsInAI

AI is no longer just about text. 🤖 Multimodal AI can work with: 👁️ Images 🎙️ Audio 📝 Text 🎬 Video The result? More natural ways to interact with intelligent systems. One AI. Multiple ways to understand the world. #MultimodalAI #AI

Multimodal AI is blowing up in 2026! It's AI that mixes text, pics, videos, and sound, not just words. Understands the world like we do. Why trending? Tools like Gemini and Grok make it real. 👇 #MultimodalAI #AI2026

Getting ready for today on our #MultimodalAi #RadAIchat this is a nice read! Shout out to @FelipeKitamura and @EricTopol! See you later today discussing multimodal AI! @khosravi_bardia @Radiology_AI

A transformer model for chest X-rays with electronic health records @radiology_rsna @DanielTruhn @FirasKhader96 @jnkath Our accompanying editorial critique w/ @FelipeKitamura

🚀 just made its biggest AI upgrade yet. GLM-4.5 now comes loaded with the multimodal power of GLM-4.5V — one model for text, images, videos, GUIs & more

5. Why it matters This is all-scenario visual reasoning — not just captions, but deep analysis for real-world workflows.

Tweet 6 — Your turn 🚀 Stop imagining what’s possible with multimodal AI — start building with it. ⚡ Try GLM-4.5V now: 🤝 Connect with the community: #GLM45V #MultimodalAI #ZaiAI

Feeling incredibly grateful to share that our team has received funding from the @NIH to continue advancing #multimodalAI in critical care. This work means a lot — and it’s just getting started.

Welcome to focus on our new paper in the field of Multimodal Hallucination: "Unified Hallucination Detection for Multimodal Large Language Models." 🤖💡 #AI #MultimodalAI #MachineLearning #NLP #LLMs #EasyDetect #Hallucination #UniHD 🔍 Our latest paper, "Unified Hallucination Detection for Multimodal Large Language Models," pioneers a unified approach, UniHD to detecting hallucinations in MLLMs (e.g., Text2Image & Image2Text) and introduces MHaluBench, a benchmark that spans diverse hallucination types and multimodal tasks. 📌ArXiv: 📌Home page: 📌Dataset: 📌Code: 📊 Benchmark: MHaluBench may be your go-to resource for various multimodal hallucination detector assessment. It is constructed by LLMs with crowdsourcing and has a balanced distribution of instances across three pivotal tasks, including 200 exemplars for Image Captioning, 200 for VQA, and an additional 220 dedicated to Text-to-Image Generation. 🛠️ Methodology: UniHD is our tool-augmented framework that systematically integrates evidence from various auxiliary tools. Here's how it works: >1️. Essential Claim Extraction: Identifies key claims in generated responses or user queries. >2. Autonomous Tool Selection for Claim: MLLMs like GPT-4/Gemini autonomously craft questions that help select the right tools for claim validation. >3. Parallel Tool Execution: A suite of specialized tools runs simultaneously, collating evidence from external knowledge to assess potential hallucinations. >4. Hallucination Verification with Rationales: Combines evidence to enable MLLMs to make informed decisions on hallucinations, providing clear explanations. 🧪 Experiments: We conduct comprehensive experiments with different MLLMs, demonstrating that MHaluBench poses a challenging benchmark for multimodal hallucination detection. GPT-4V surpasses Gemini as the detector base, and UniHD empowered by GPT-4V shows superior detection across the board. We also notice that UniHD, powered by GPT-4V, consistently excels, aligning with top leaderboards and underscoring its effectiveness for evaluating hallucinations in MLLMs. 🌟UNIHD + GPT-4V = The great combination for detecting hallucinations in the latest MLLMs, offering a reliable measure for hallucination rankings. 🔄 Our work is still in progress. Welcome to follow and provide valuable feedback.