HKUST and ByteDance introduce AdaMM multimodal memory

// 46d agoRESEARCH PAPER

HKUST and ByteDance introduce AdaMM multimodal memory

Developed by researchers from HKUST and ByteDance, AdaMM is a multimodal memory framework that integrates structured analytic memory alongside vector retrieval. This hybrid approach enables AI agents to perform structured queries, temporal aggregation, filtering, and comparative analysis across long-term interaction histories.

// ANALYSIS

Vector retrieval alone is insufficient for long-term agent interaction histories that demand temporal reasoning and structured data aggregation.

–Combines classic vector retrieval with structured analytic memory for comprehensive multimodal history management.
–Enables AI agents to execute temporal comparisons, structured filtering, and multi-turn data aggregation.
–Facilitates powerful local memory scaffolds that reduce reliance on massive, costly cloud LLM context windows.

// TAGS

adammai-agentsmultimodal-memoryhkustbytedanceanalytic-memorymachine-learning

DISCOVERED

46d ago

2026-08-05

PUBLISHED

46d ago

2026-08-05

RELEVANCE

8/ 10

AUTHOR

Discover AI

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

NEWS2h ago

GPT-6 Astra captures 13% OpenRouter enterprise spend

OpenRouter data reveals that OpenAI's GPT-6 Astra captured roughly 13% of enterprise AI spend within two weeks of release, demonstrating rapid commercial adoption for frontier agentic models. Concurrently, OpenAI classified Astra at the Critical cybersecurity threshold under its Preparedness Framework—the first model to reach this rating—due to demonstrated capabilities in autonomously discovering zero-day vulnerabilities and generating functional exploits against hardened systems without human intervention.

OPEN SOURCE4h ago

DIFF launches LLM-ready DayZ script browser

DIFF (DayZ Internal File Finder) is an open-source documentation explorer and reference tool built for the DayZ modding and scripting community. Automatically generated from Bohemia Interactive's script repository across game builds, the platform catalogs thousands of Enforce Script classes, methods, fields, globals, and file structures. It enables developers to search the codebase, trace changes between updates, and contribute crowd-sourced annotations via a shared notes repository. Crucially, DIFF is designed with modern AI workflows in mind, offering structured machine-readable endpoints (`llms.txt`, `api.json`), agent integration guides, and one-click Markdown exports for feeding context directly into LLMs.

OPEN SOURCE4h ago

Google releases AX distributed agent runtime

Developed by Google, AX (Agent Executor) is an open-source distributed agent runtime built to solve the reliability, sandboxing, and scaling challenges of long-running AI agent workloads. Instead of competing directly with cognitive orchestration frameworks like LangGraph or CrewAI, AX serves as the underlying execution engine that manages containerized sandboxes, workspace environments (including Git repositories and MCP servers), and network egress fencing through declarative YAML definitions.