Author

Amap Tech

Official Amap technology account showcasing all of Amap's technical innovations.

261

Articles

Likes

1.6k

Views

Comments

Latest from Amap Tech

100 recent articles max

Amap Tech

Jun 25, 2026 · Artificial Intelligence

ReaGeo: The First End‑to‑End LLM Geocoding Framework Linking Precise Mapping and Spatial Correlation

ReaGeo, a novel end‑to‑end geocoding system built on the Qwen2.5‑3B large language model, converts address text directly into Geohash sequences using chain‑of‑thought reasoning and GRPO reinforcement learning, achieving an average error of 119.6 m and 97.2 % accuracy within 500 m on Beijing data, surpassing commercial APIs and academic baselines while also modeling broader spatial correlation for line‑ and area‑type queries.

GeoHashGeocodingLLM

0 likes · 15 min read

ReaGeo: The First End‑to‑End LLM Geocoding Framework Linking Precise Mapping and Spatial Correlation

Amap Tech

Jun 24, 2026 · R&D Management

Boosting Cross-Stack Development with Harness-Loop: AI-Powered Full-Stack Delivery

The article introduces the harness-loop engineering framework, which equips AI with a structured workspace and iterative loop to identify impact scope, orchestrate cross-repository tasks, perform differentiated verification, and capture experience, thereby streamlining the full-stack development of Gaode’s Street-Ranking feature across backend, middleware, and client layers.

AIautomationcross-stack

0 likes · 19 min read

Boosting Cross-Stack Development with Harness-Loop: AI-Powered Full-Stack Delivery

Amap Tech

Jun 24, 2026 · Artificial Intelligence

From Workflow to HermesAgent: Language‑Driven Image Selection for Gaode Street Rankings

The article details how Gaode transformed its massive POI image‑selection pipeline from a fragile, SQL‑driven workflow to a hybrid deterministic‑plus‑Agent architecture called HermesAgent, achieving a 48‑fold speedup, multimodal quality checks, a knowledge base, and natural‑language‑driven production while maintaining stability and observability.

AI-powered image selectionAgent-based rankingHermesAgent

0 likes · 23 min read

From Workflow to HermesAgent: Language‑Driven Image Selection for Gaode Street Rankings

Amap Tech

Jun 23, 2026 · Artificial Intelligence

GrowLoop: Turning Subjective Dialogue Quality into a Rational Benchmark

GrowLoop proposes a self‑evolving loop that uses a few human seed annotations and large‑language‑model meta‑reflection to automatically generate and refine scoring rubrics and test questions for open‑domain dialogue, enabling reliable benchmarking where no fixed standard exists.

LLMReward Modelingbenchmark

0 likes · 23 min read

GrowLoop: Turning Subjective Dialogue Quality into a Rational Benchmark

Amap Tech

Jun 22, 2026 · Artificial Intelligence

Three Amap Papers Accepted at IROS 2026: VLN Navigation, VLA Latency Correction, and Diffusion‑Based Quadruped Control

IROS 2026 received 4,348 submissions and accepted 1,585 papers (36% acceptance); Amap had three papers selected, covering online semantic‑affordance navigation, an asynchronous edge adapter for VLA‑based navigation, and a diffusion‑guided constraint‑aware framework for high‑fidelity quadruped locomotion.

asynchronous VLA navigationdiffusion modelsembodied AI

0 likes · 8 min read

Three Amap Papers Accepted at IROS 2026: VLN Navigation, VLA Latency Correction, and Diffusion‑Based Quadruped Control

Amap Tech

Jun 18, 2026 · Artificial Intelligence

MapSatisfyBench: From Task Completion to Decision Satisfaction for Map Agents

MapSatisfyBench, a new benchmark co‑developed by Amap and Peking University, shifts map‑agent evaluation from simple task‑completion to quantifiable decision‑satisfaction by extracting implicit user demands via a restore‑identify‑filter pipeline, constructing a five‑dimensional truth schema and revealing that current LLM agents often complete tasks yet fail to meet hidden user preferences.

Decision SatisfactionFive-Dimensional MetricsImplicit Demand

0 likes · 14 min read

MapSatisfyBench: From Task Completion to Decision Satisfaction for Map Agents

Amap Tech

Apr 19, 2026 · Artificial Intelligence

From Pixels to the Physical World: Inside Gaode’s ABot Full‑Stack

Gaode leverages 20 years of spatiotemporal data to unveil a three‑layer ABot stack—World Model, navigation (N series), manipulation (M series) and a Harness architecture—that embeds physical laws, self‑evolves through a dual data‑training engine, and achieves benchmark‑leading performance across wheeled, quadruped and humanoid robots.

benchmarkembodied AImanipulation

0 likes · 16 min read

From Pixels to the Physical World: Inside Gaode’s ABot Full‑Stack

Amap Tech

Apr 1, 2026 · Artificial Intelligence

Can World Models Truly Understand Interaction? Inside the Omni-WorldBench

Omni-WorldBench introduces a comprehensive benchmark that shifts world‑model evaluation from visual fidelity to interactive response, detailing its two‑part suite, metric design, extensive prompt taxonomy, and experimental results that reveal current models' strengths and limitations in causal and temporal reasoning.

AIOmni-WorldBenchbenchmark

0 likes · 11 min read

Can World Models Truly Understand Interaction? Inside the Omni-WorldBench

Amap Tech

Mar 30, 2026 · Artificial Intelligence

ABot-M0: A Unified VLA Framework Solving the One‑Brain Many‑Forms Robotics Challenge

ABot-M0 is an open‑source Vision‑Language‑Action foundation model that unifies fragmented robot data, introduces Action Manifold Learning for smoother action prediction, and offers a plug‑and‑play dual‑stream perception architecture, achieving state‑of‑the‑art results on major manipulation benchmarks.

action manifold learningembodied AIfoundation model

0 likes · 4 min read

ABot-M0: A Unified VLA Framework Solving the One‑Brain Many‑Forms Robotics Challenge

Amap Tech

Mar 20, 2026 · Artificial Intelligence

How ABot-PhysWorld Achieves Physical Consistency in Embodied Video Generation

ABot-PhysWorld introduces a physically consistent video generation framework for embodied AI, leveraging the PAI‑Bench benchmark, large‑scale multi‑modal data, DPO preference alignment, and dense action maps to surpass SOTA models in both visual quality and physical plausibility across diverse robotic tasks.

Physical Consistencybenchmarkdeep learning

0 likes · 15 min read

How ABot-PhysWorld Achieves Physical Consistency in Embodied Video Generation