Termsslug: multimodal-perception

Multimodal Perception

多模态感知

The ability of models to process and understand inputs from multiple modalities like images and text.

Curated Intelligence

10 items · Page 1/1
InfoQ 中文T2·Industry
58 pts
JD Logistics Unveils Super Brain 3.0 with Second-Level Route Planning

JD Logistics has upgraded its Super Brain model to version 3.0, comprising five industrial-grade models for forecasting, decision-making, multimodal perception, spatiotemporal reasoning and embodied control. The company claims end-to-end optimal route planning for hundreds of millions of parcels now solves in seconds rather than minutes, and says the stack will open to manufacturing and retail industries.

IT之家 · 智能时代T2·Industry
42 pts
Amazon Exec Panos Panay Envisions Ambient Computing: AI Wearables May End App Dependency

Amazon SVP Panos Panay argued at TechCrunch Disrupt that the future of personal AI lies in ambient computing devices that proactively sense user needs, rather than traditional apps. Coinciding with the launch of the new Bee AI wearable, he emphasized privacy control as a core design principle for background-running devices and predicted an intelligent layer connecting multiple form factors instead of replacing smartphones.

TechCrunch AIT2·Industry
48 pts
Amazon's Panos Panay Envisions Ambient AI Wearables Replacing Traditional App Interactions

Amazon SVP Panos Panay outlined his 'Ambient Computing' strategy at TechCrunch Disrupt, arguing that AI wearables like the new Bee devices should use sensors and background intelligence to proactively understand user needs, moving away from passive app-based interactions. While noting that no single form factor will completely replace others, he emphasized that a unified intelligence layer connecting these devices must be highly useful, signaling a shift in how tech giants define personal AI interfaces.

InfoQ 中文T2·Agents
68 pts
Alibaba Qwen Launches Screenless AI Glasses N1: 44g Lightweight Design with Eye Tracking Focuses on Voice Interaction for Real-World Context

Alibaba Qwen launched the screenless AI glasses N1/N1 Pro, reducing weight to 44g by removing displays and prioritizing voice interaction to address comfort issues. Positioned as Personal Agent hardware, it integrates a 50MP sensor, eye tracking, and biometric sensing to provide AI with real-world multimodal context. Despite early challenges like network dependency and rough edge-side parsing, this 'subtraction' design aims to break the industry cycle of feature bloat and low user retention.

IT之家 · 智能时代T2·Tools
42 pts
XGIMI Launches MemoMind One AI Glasses: Focus on Privacy-First Design and All-Day Smart Memory

XGIMI has launched the MemoMind One AI glasses, featuring a camera-free design for privacy and core capabilities in all-day AI memory and automated meeting minutes. Priced at 2699 RMB, it supports binocular display, real-time translation in 53 languages, and navigation prompts, aiming to streamline fragmented information capture via voice interaction.

The Verge AIT2·Tools
42 pts
Ray-Ban Meta Audio Glasses Review: Removing Camera Reduces Privacy Concerns but Limits Multimodal AI Utility

Victoria Song reviews the camera-less Ray-Ban Meta Audio glasses, noting that while removing the lens alleviates social privacy concerns, it significantly hampers multimodal context awareness. The piece highlights persistent friction in basic voice-assistant tasks like calendar syncing and criticizes the cumbersome onboarding agreement process.

InfoQ 中文T2·Agents
68 pts
Qwen Launches N1 Pro AI Glasses and Bose Earbuds: Integrating Iris Payment and Biometric Monitoring to Extend Personal Agents

On October 13, Qwen launched the N1/N1 Pro AI glasses and co-branded Bose AI earbuds, extending Personal Agent capabilities to wearables. The N1 Pro features industry-first continuous biometric monitoring (heart rate, SpO2) alongside eye-tracking interaction and iris-based payment with a false acceptance rate below one-in-a-million. The earbuds support real-time translation across 89 languages and background task delegation, aiming for proactive service through all-day multimodal perception.

InfoQ 中文T2·Agents
78 pts
OpenAI DevDay Deep Dive: Dot Agent Enables Autonomous Cloud Linux Operations; Decisions API Prototype Built in One Week

OpenAI launched the Dot agent and computer use capabilities at DevDay, enabling autonomous execution of browser and desktop tasks within cloud Linux environments via multimodal perception (screenshots, accessibility APIs, code). Key breakthroughs include significantly improved error handling and retry mechanisms, alongside the new Decisions API inspired by Jev. The latter leverages Luna weights with optimized inference for millisecond-level structured decision returns. However, human oversight boundaries for critical actions like payments remain undefined.

Kimi 博客 (月之暗面)T1·Models
25 pts
Kimi Launches PerceptionBench for AI Models

Moonshot AI (Kimi) has launched PerceptionBench, a new benchmark designed to evaluate the multimodal perception capabilities of large models. As the provided text consists primarily of navigation and footer links without specific details on methodology, dataset composition, or initial results, only the announcement itself is confirmed.