AI Briefing#multimodal-ai
The latest AI and developer news about #multimodal-ai, with the original source and a short summary.
The latest AI and developer news about #multimodal-ai, with the original source and a short summary.
MindTopo measured the gap between VLMs' static topological recognition and interaction planning capabilities.
Google's Gemini app has surpassed 1 billion monthly users.
Omnichannel AI agents maintain conversation context across channels to provide a consistent customer experience.
Google was named a Leader in Forrester's AI Platforms evaluation, receiving the highest score in the Strategy category.
JoyAI has released a 16B model that causally edits live video streams in real time based on natural language instructions.
MiniMax has released H3, an open-weight model that generates video and stereo audio simultaneously.
The Qwen development team teased the launch of a 27B model and key specifications for upcoming models.
SimpleMem compresses long-term memory to boost LoCoMo performance while cutting inference tokens by up to 30x.
Opus 5 implemented a Lord of the Rings scene in code in just 2 hours.
A roundup of 10 AI/ML papers covering real-world agent evaluation and token efficiency.
Seedance 2.5 has launched with enhanced 30-second generation and multimodal reference and precise editing capabilities.
Microsoft has unveiled a new MAI model lineup including a high-performance image generation model and a high-speed voice model.
A human-annotated LVSum benchmark has been released to evaluate temporal accuracy in long video summarization.
LG AI Research has revealed the technical principles of EXAONE, a multimodal model that understands text and images bidirectionally.
LG AI Research presented research achievements related to continual learning and scale-invariant networks at ECCV 2022.
We introduce ChartInstruct, a general-purpose chart vision-language model that applies Instruction Tuning to maximize chart understanding and reasoning capabilities.
Thinking Machines has unveiled Inkling, a 975B-parameter multimodal open-weight model.
Moonshot AI's Kimi K3 has become available on Vercel AI Gateway.
Google announced Gemma 4 E2B, an on-device AI model that runs natively on the Tensor G5 TPU of the Pixel 10 series.
DoorDash built a system that automatically generates large-scale food metadata using LLM Jury and context-optimization agents.
ByteDance has launched Seedream 5.0 Pro, a multimodal image generation model optimized for design workflows.
A hybrid architecture called DCG is proposed, combining neural networks and symbolic knowledge structures to help AI maintain concepts continuously.
Google unveiled a more integrated AI ecosystem through the launch of Gemma 4 12B and the Android 17 update.
It was revealed that the hallucination phenomenon in which the multimodal model Inter-1 repeatedly outputs a specific phrase on silent videos is caused by a combination of prompt examples and post-training.