読書リサーチレーダー投資フレームワーク
ログイン / 新規登録
すべて技術製品投資株式市場その他

読書

2026-11-1392 投稿

更新

How the A2A Protocol Enables Cross-Organizational Collaboration for Agents Across Frameworks: From Agent Card Discovery to Bidirectional Callbacks

トピック · A2A代理协议重大
2026-11-12 23:42 GMT+8

Cross-organizational multi-agent systems face a specific constraint: two institutions build and run their own agents, which need to discover each other and exchange tasks without exposing internal implementations or requiring both sides to use the same framework. Without a standardized protocol, every new customer or framework combination requires an ISV to write custom integration code. The A2A protocol attempts to replace these disparate paths with a single interface.

The reference architecture provided in the AWS Machine Learning blog runs three agents across two independent AWS accounts. The enterprise-side orchestrator is based on LangChain, invoking Anthropic Claude Sonnet 4.6 for reasoning routing; the SaaS-side compliance agent is built on the Strands Agents SDK and deployed on Amazon Bedrock AgentCore. Both use completely different frameworks and SDKs but communicate via A2A’s JSON-RPC over HTTP. The orchestrator does not depend on the Strands SDK, nor does the compliance agent depend on LangChain—this is the practical meaning of protocol-level interoperability: as long as both sides speak A2A and publish compatible Agent Cards, the frameworks can be swapped independently.

Agent Cards are the core discovery layer of this mechanism. They publish a JSON document at the /.well-known/agent-card.json path, describing the agent’s name, capabilities, and skills. The orchestrator fetches this card at startup, creating a LangChain tool for each skill listed on the card, after which the large language model decides which tool to call based on the user’s question. This means that when an ISV adds or modifies capabilities, it only needs to update the Agent Card; the orchestrator automatically acquires the new skills upon its next restart, eliminating the need for the two teams to coordinate code changes.

A key design decision lies in what the card publishes. Although the compliance agent has three internal tools (listing accounts, executing scans, and retrieving remediation steps), the Agent Card publishes only one curated skill: 'Scan your AWS account, returning findings and remediation steps by severity.' The calling model sees what the product does, not how it is built. This abstraction allows the ISV to refactor its internal architecture—swapping models, tools, or implementations—as long as the skills described in the card remain unchanged, leaving customer integrations unaffected.

Bidirectional callbacks are what set this solution apart from typical agent demonstrations. Most inter-agent communication demos show a unidirectional request-response pattern: a client calls a server agent and waits for the result. However, this pattern fails when the called agent needs information that only the caller possesses. In this architecture, the compliance agent requires enterprise context before scanning—such as which accounts are production environments, which internal policies apply, and who to escalate to—information known only to the enterprise side. The compliance agent calls back to the enterprise context agent via A2A, obtaining a token using its own account’s identity provider and sending a message to the callback address registered by the customer during onboarding. This makes the two agents true peers rather than a client and a server.

ソース:Amazon Web Services

リサーチ

a16z-backed startup Phota Labs launches AI photo editor that preserves user identity

トピック · Phota Labsクイックテイク
2026-11-13 01:00 GMT+8

Phota Labs, a startup backed by Andreessen Horowitz (a16z), launched its iOS application on November 12, 2026. The app serves as a personal AI photo assistant designed to transform everyday snapshots into images that resemble professional photography. Unlike traditional filters or manual retouching tools, Phota Labs emphasizes "identity preservation," ensuring that adjustments to lighting, composition, and lens effects do not distort or replace the natural appearance of the people in the photo.

The company was founded by Cecilia Zhang and Zach Xia, who previously worked together at Adobe on computational photography products including Project Indigo, Lightroom, and Photoshop. They identified a gap where existing generative AI tools often altered a person's appearance when fixing blurry or low-quality images. Phota Labs addresses this by building a proprietary personalization model. "For Phota, we built our own tech," Zhang told TechCrunch. "We built this personalization technology that actually learns about people’s identity and is able to preserve them in any transformation we do to a photo."

The app is launching as a free service to build an initial user base before exploring a freemium model, with plans for an Android version in the future. Users can select various photography styles ranging from subtle edits to dramatic transformations like fisheye lenses or alternative angles. All images edited with Phota Labs feature invisible watermarks to identify them as AI-altered content when shared on social platforms.

ソース:TechCrunch

リサーチ

SoftBank Reports $7.1 Billion Quarterly Loss Driven by Gas Turbine Deposit Write-Downs

トピック · 软银燃气轮机减值重大
2026-11-13 01:00 GMT+8

Japanese investment giant SoftBank Group Corp. disclosed a net quarterly loss of $7.1 billion in its earnings report released Tuesday. According to TechCrunch, more than half of this loss stemmed from write-downs related to natural gas power generation equipment ordered for artificial intelligence data centers.

Starting about two years ago, SoftBank, through its subsidiary SB Energy, began placing massive orders for gas turbines. Reports indicate it paid non-refundable deposits for over 13 gigawatts of capacity, nearly 50% more than it actually needed. This strategy aimed to secure scarce turbine supplies against the surging power demands driven by AI compute growth. However, as market supply-demand dynamics shifted and actual energy consumption expectations for AI data centers adjusted, SoftBank deemed much of these prepayments worthless, leading to the significant impairment.

Peter Gardett, CEO of Noreva, noted that this loss reflects a process of "price discovery": SoftBank overpaid and over-ordered to secure turbines quickly. While AI's demand for electricity remains substantial, SoftBank's setback may signal a ceiling in the industry's blind expansion into natural gas power. Brent Nelson, Senior Managing Director at Ascend Analytics, stated that as operators better understand actual data center running requirements, grid connections paired with large battery storage might become more cost-effective choices, potentially slowing the expected development pace of the natural gas power market.

ソース:TechCrunch

リサーチ

Rockwell Automation Guides for 6-9% Organic Growth in FY2027 Amid Record Orders

トピック · 罗克韦尔自动化增长预期重大
2026-11-12 13:12 GMT+8

Rockwell Automation (NYSE: ROK), a leading US industrial automation company, expects organic growth of 6% to 9% in fiscal 2027, according to Chairman and CEO Blake Moret speaking at a Baird investor conference. The outlook is supported by strong demand for automation equipment, higher prices, and potential acquisitions.

Moret noted that demand from data centers, semiconductors, automotive, and life sciences exceeded expectations in fiscal 2026. The automotive sector, accounting for about 10% of revenue, strengthened particularly in the latter part of the year. Pharmaceutical demand was partly supported by GLP-1 drug production, though Moret described the strength across life sciences as broad-based.

Order momentum underpins this positive outlook. Moret reported that fiscal 2026 orders for new US physical plant capacity were the highest in the company's history, and the pipeline entering fiscal 2027 was larger than a year earlier. Fourth-quarter orders grew at a mid-teens rate, with the first quarter starting well.

Regarding pricing, the fiscal 2027 guidance includes three percentage points of price contribution but excludes a planned December price increase. Moret identified memory cost inflation as a key driver for higher prices, although the company has largely managed supply shortages. Logix control products saw 25% growth in Q4, attributed to underlying demand and pricing rather than customer restocking.

Software and AI strategy also contribute to growth. The company targets 7% growth in annual recurring revenue (ARR) in fiscal 2027 and plans to launch FactoryTalk Conductor, a new software offering using agents to speed up control system integration. Analyst Quinn Fredrickson noted the guidance implies about 80 basis points of enterprise operating margin expansion. Moret added that Rockwell expects inorganic growth to return, prioritizing profitable hardware businesses that can increase share in Europe or Asia and fit its US distribution network.

ソース:Yahoo Finance

リサーチ

Microsoft VP Scott Hanselman Releases Open-Source Tool GPU Taskman for Monitoring VRAM and Local LLM Memory Needs

トピック · GPU Taskmanクイックテイク
2026-11-12 23:37 GMT+8

Scott Hanselman, Microsoft Vice President and software developer, has released GPU Taskman, an open-source task manager specifically designed to monitor graphics processing unit (GPU) video memory usage. The tool runs on Windows, Linux, and macOS, combining process lists with device data such as VRAM occupancy, load, and temperature via a terminal interface to help developers understand which processes are consuming memory.

For locally running Large Language Models (LLMs), GPU Taskman identifies loaded runtime environments and model files, estimating the storage required for weights and the Key-Value Cache (KV-Cache). The KV-Cache is a critical data structure in Transformer architectures used to accelerate inference, growing in size with context length. The tool distinguishes between measured data and estimates, avoiding the misidentification of unused disk models as active processes.

The latest version, v0.4.0, adds support for additional Intel GPU metrics on Linux and device monitoring for Huawei Ascend NPUs. Developers explicitly state that the software is in an early experimental stage, noting that model attribution and memory allocation calculations are not direct measurements, advising users to treat estimated data with caution.

ソース:iX Magazin

リサーチ

Agent DLC: When Retrieval Precision is Perfect but the Agent Can't Answer, How Should the Evaluation Framework Be Designed?

トピック · Agent DLC重大
2026-11-12 23:38 GMT+8

An internal HR policy assistant answers three questions: performance reviews, benefits enrollment, and sick-leave certificates. Chunk-level retrieval precision—which measures whether each retrieved document block as a whole is relevant to the question—scores a perfect 1.00 for all three questions. If the release threshold relies solely on this metric, all three answers would pass.

However, when each document block is broken down into individual facts and their necessity is checked one by one, the fact-level precision scores are 0.59, 0.53, and 0.09 respectively. The third question indicates that roughly nine out of ten specific facts in the retrieved material are irrelevant to what was asked. Chunk-level precision only asks 'is this passage on topic'; it does not measure 'how much of this passage is actually useful.'

In a blog post, the AWS engineering team uses this measured case to introduce the core argument of the Agent Development Lifecycle (Agent DLC) framework: traditional unit testing and CI/CD remain necessary for agents, but they are insufficient due to three characteristics. First, agents are non-deterministic; identical inputs can produce different outputs and execution paths. Second, changes to prompts, tool descriptions, retrieval settings, or memory configuration can alter behavior in ways undetectable by text review alone. Third, model endpoints, knowledge bases, and external APIs can change independently of the code repository and affect agent behavior.

The framework organizes development activities into two nested loops. The inner loop is: score → inspect the path → change one variable → re-evaluate, improving the agent against fixed standards. The outer loop is: define → build → evaluate → release → observe → feedback, allowing the standards themselves to change. The authors provide an example prone to confusion: if production users ask the agent to look up information on behalf of a family member, but no standard covers this scenario, this is an outer-loop problem—it requires defining authorization boundaries and adding examples, then using the inner loop to meet the updated standard. Confusing the two leads teams to attempt filling a missing standard with prompt modifications.

Returning to the sick-leave certificate case. After modifying the prompt, the performance review question improved from 0.83 to 0.87, and fact-level precision jumped from 0.59 to 0.85; however, the groundedness of the sick-leave certificate answer dropped from 0.39 to 0.31, and fact-level precision fell from 0.09 to 0.04—one change yielding two opposite results. Diagnosis points to a knowledge coverage gap: there were no dedicated sick-leave policy documents in the corpus, only brief mentions within the leave-management policy. Telling the model to 'answer strictly based on retrieved content' cannot conjure missing information out of thin air.

Evaluation criteria are organized into five dimensions—domain understanding, quality, responsibility, cost, performance—and three tiers: hard constraints (boolean values, violations block release), release gates (pass-rate thresholds over k runs), and monitoring only (tracked but do not block). The authors highlight a specific case: a benefits enrollment answer passed with a groundedness of 0.90, yet contained the sentence 'I noticed your profile shows your preferred contact method is social media'—a profile that does not exist. The model combined an option from the policy document with an unrelated answer from a public FAQ, packaging it as a fact about the user. A high groundedness score failed to capture this false personal claim.

ソース:Amazon Web Services

リサーチ

Canadian E-commerce Platform Shopify Deploys 'Cyber Chaos Monkey': AI Agents Autonomously Attack Production Infrastructure to Find Bugs

重大
2026-11-12 22:57 GMT+8

Canadian e-commerce platform Shopify disclosed on November 12, 2026, via its engineering blog, an internal security system named "Cyber Chaos Monkey" (CCM). The system leverages frontier AI models combined with Shopify's deep knowledge of its software architecture, source code, and infrastructure to proactively launch simulated attacks on its production environment, aiming to identify and fix vulnerabilities before external attackers can exploit them.

The core logic of CCM relies on information asymmetry. Shopify noted that as frontier models drive the cost of vulnerability discovery towards zero, reports received through its Bug Bounty program have increased by nearly 300% since 2024, making manual review and remediation increasingly difficult. CCM addresses this by providing AI agents with live test credentials and comprehensive access to the system, enabling them to operate like advanced bug bounty hunters but at a scale far beyond human capability.

The system architecture is divided into two components: "Discovery" and "Sandbox." Discovery receives hypotheses or leads and uses the latest cyber models to refine potential vulnerability locations. Sandbox executes specific exploitation attempts, including installing packages, building fuzzers, and sending direct requests to Shopify's production environment while analyzing logs. This closed loop allows AI agents to iteratively optimize their attack strategies based on feedback.

To ensure safety, Shopify implemented multi-layered controls. Technically, all operations occur within containerized environments following the principle of least privilege, with network access restricted via credential proxies and strict egress allowlists. Governance-wise, a "Two-Person Integrity" (TPI) mechanism was introduced, using a separately aligned AI agent to monitor the execution agent's behavior logs in real time. If the monitoring agent detects high-risk behavior violating the "Rules of Engagement," it can immediately terminate the task.

According to Shopify, CCM has executed tens of thousands of offensive engagements, identified and acted on dozens of impactful findings, and caused no outages or incidents. This practice marks a shift in enterprise security defense from passive response to proactive, automated red-teaming driven by AI.

ソース:Shopify

リサーチ

Graph Attention Networks: Solving Inductive Learning Challenges in Graph-Structured Data with Self-Attention Mechanisms

検証済み 2026-11-13 22:12 GMT+8

長期読解 · 《Graph Attention Networks》(2017)

In the field of deep learning, Convolutional Neural Networks (CNNs) have achieved great success in processing grid-structured data such as images. However, many real-world datasets, such as social networks, biological networks, or communication networks, do not possess regular grid structures but instead exist in graph form. Early graph neural network methods were primarily divided into spectral and non-spectral approaches. While effective, spectral methods typically rely on eigendecomposition of the Laplacian operator, meaning that trained models cannot be directly applied to new graphs with different structures, limiting their flexibility in inductive learning tasks.

To address this challenge, Petar Veličković et al. proposed Graph Attention Networks (GATs). The core innovation of this architecture is the introduction of masked self-attention layers. Unlike traditional methods that depend on fixed neighborhood sizes or predefined graph structures, GAT allows nodes to "attend" to features of their neighbors and implicitly assign different weights to each neighbor. This mechanism not only avoids expensive matrix operations (such as inversion) but also does not require prior knowledge of the complete graph structure. By stacking these attention layers, the model can efficiently process node pairs in parallel and flexibly adapt to nodes with varying degrees.

A key advantage of GAT is its inductive learning capability. Since attention weights are dynamically computed based on node features rather than fixed graph topology, the model can be directly applied to entirely new graph structures unseen during training. The paper notes that this approach simultaneously addresses several critical challenges faced by spectral-based graph neural networks and achieves state-of-the-art or comparable results on citation network datasets such as Cora, Citeseer, and Pubmed, as well as protein interaction datasets. For researchers interested in understanding how to extend Transformer-based attention mechanisms to non-Euclidean data, this is a foundational seminal work. It is recommended to read "Graph Attention Networks" (submitted 2017, published at ICLR 2018), focusing on how it solves inductive learning problems in graph-structured data through self-attention mechanisms.

ソース:《Graph Attention Networks》(2017)

リサーチ

このビューで最新の記事まで読みました

ログイン / 新規登録
読書リサーチレーダー投資フレームワーク

Regarding authentication, each account has its own Amazon Cognito user pool, with both sides obtaining tokens from their respective pools. The receiving side trusts only tokens issued by the other party’s pool. During onboarding, the parties exchange only public values—identity provider URLs, client IDs, scopes, and endpoint URLs—without sharing any secrets. Each A2A task request carries a JWT, which the receiver validates for issuer, client, and scope. The only endpoint that does not require authentication is the context agent’s Agent Card, as it only publishes descriptive metadata.

End-to-end testing validated the entire chain: the orchestrator discovered the compliance agent, routed the request, the compliance agent called back to the context agent to retrieve account metadata, executed the scan, and returned a rich report containing data from both agents. The test passed if the reply included compliance findings, contained an on-call rotation code known only to the context agent, and the partner API access logs recorded a 200 response for the callback client. Subsequently, seven trust boundary checks confirmed that each token was valid only where trusted, rejecting errors, missing, or forged tokens.

The applicability boundaries of this reference implementation are noteworthy. It uses synthetic data to simulate compliance findings and does not demonstrate latency, throughput, or fault recovery performance in a production environment. The bidirectional callback pattern requires both parties to maintain identity provider configurations and callback registrations, resulting in higher operational complexity than unidirectional calls. The blog itself lists productionization recommendations: splitting runtimes per customer, replacing client secrets with key-based authentication, enabling PrivateLink and KMS encryption, and configuring VPC flow logs. These notes indicate that the current architecture is a pattern validation, not a ready-to-deploy production product.

The practical implication for readers is this: if an organization is building agent systems that require collaboration across teams or companies, A2A’s Agent Card discovery combined with bidirectional OAuth callbacks offers an alternative to custom integrations. It does not require both parties to use the same framework, share secrets, or modify caller code to expand capabilities; instead, capability expansion is achieved by updating the card. However, the prerequisite for this path is that both parties accept protocol constraints and maintain their own identity infrastructure—an operational commitment in itself.

The framework's practical advice is: select a clearly scoped business task and establish an evaluation baseline within a week. On days two, business owners and engineers agree on acceptance criteria, tiers, and thresholds, preparing 20 to 30 golden examples. From day three to five, implement the evaluators and run the golden set repeatedly; the published baseline includes failure cases. In the second week, change one variable and re-evaluate under the same conditions, recording the release decision and its evidence. The authors emphasize that scores may fluctuate between runs, borderline passes can become failures, and differences should be interpreted alongside repeated runs and underlying evidence.