MAROKO133 Update ai: Why Google’s File Search could displace DIY RAG stacks in the enterpr

📌 MAROKO133 Hot ai: Why Google’s File Search could displace DIY RAG stacks in the

By now, enterprises understand that retrieval augmented generation (RAG) allows applications and agents to find the best, most grounded information for queries. However, typical RAG setups could be an engineering challenge and also exhibit undesirable traits. 

To help solve this, Google released the File Search Tool on the Gemini API, a fully managed RAG system “that abstracts away the retrieval pipeline.” File Search removes much of the tool and application-gathering involved in setting up RAG pipelines, so engineers don’t need to stitch together things like storage solutions and embedding creators.  

This tool competes directly with enterprise RAG products from OpenAI, AWS and Microsoft, which also aim to simplify RAG architecture. Google, though, claims its offering requires less orchestration and is more standalone. 

“File Search provides a simple, integrated and scalable way to ground Gemini with your data, delivering responses that are more accurate, relevant and verifiable,” Google said in a blog post. 

Enterprises can access some features of File Search, such as storage and embedding generation, for free at query time. Users will begin paying for embeddings when these files are indexed at a fixed rate of $0.15 per 1 million tokens. 

Google’s Gemini Embedding model, which eventually became the top embedding model on the Massive Text Embedding Benchmark, powers File Search. 

File Search and integrated experiences 

Google said File Search works “by handling the complexities of RAG for you.” 

File Search manages file storage, chunking strategies and embeddings. Developers can invoke File Search within the existing generateContent API, which Google said makes the tool easier to adopt. 

File Search uses vector search to “understand the meaning and context of a user’s query.” Ideally, it will find the relevant information to answer a query from documents, even if the prompt contains inexact words. 

The feature has built-in citations that point to the specific parts of a document it used to generate answers, and also supports a variety of file formats. These include PDF, Docx, txt, JSON and “many common programming language file types," Google says.

Continuous RAG experimentation 

Enterprises may have already begun building out a RAG pipeline as they lay the groundwork for their AI agents to actually tap the correct data and make informed decisions. 

Because RAG represents a key part of how enterprises maintain accuracy and tap into insights about their business, organizations must quickly have visibility into this pipeline. RAG can be an engineering pain because orchestrating multiple tools together can become complicated. 

Building “traditional” RAG pipelines means organizations must assemble and fine-tune a file ingestion and parsing program, including chunking, embedding generation and updates. They must then contract a vector database like Pinecone, determine its retrieval logic, and fit it all within a model’s context window. Additionally, they can, if desired, add source citations. 

File Search aims to streamline all of that, although competitor platforms offer similar features. OpenAI’s Assistants API allows developers to utilize a file search feature, guiding an agent to relevant documents for responses. AWS’s Bedrock unveiled a data automation managed service in December. 

While File Search stands similarly to these other platforms, Google’s offering abstracts all, rather than just some, elements of the RAG pipeline creation. 

Phaser Studio, the creator of AI-driven game generation platform Beam, said in Google’s blog that it used File Search to sift through its library of 3,000 files.

“File Search allows us to instantly surface the right material, whether that’s a code snippet for bullet patterns, genre templates or architectural guidance from our Phaser ‘brain’ corpus,” said Phaser CTO Richard Davey. “The result is ideas that once took days to prototype now become playable in minutes.”

Since the announcement, several users expressed interest in using the feature.

🔗 Sumber: venturebeat.com

📌 MAROKO133 Update ai: World’s first AI firefighting system extinguishes oil fires

The Korea Institute of Machinery and Materials (KIMM) has developed a next-generation autonomous fire suppression system that can detect and extinguish oil fires aboard naval vessels even under rough sea conditions.

The AI-driven system independently verifies the authenticity of a fire, activates only when one is confirmed, and directs its suppression precisely at the source, much like a human firefighter.

The system, developed by Senior Researcher Hyuk Lee and his team at KIMM’s AX Convergence Research Center, completed successful trials aboard a real naval vessel.

It is an advanced version of the team’s earlier autonomous firefighting research, now adapted for the oil fires most common on naval ships.

Unlike traditional firefighting systems that release extinguishing agents throughout an entire compartment, KIMM’s technology targets only the fire source.

This prevents unnecessary damage during false alarms.

By using AI-based detection and reinforcement learning, the system adapts to ship movement and sea conditions to ensure accurate discharge.

The technology includes sensors, fire monitors, and a control unit with AI-based fire verification and location estimation capabilities.

It achieved a fire detection accuracy rate of more than 98 percent and can discharge foam up to 24 meters. Tests also confirmed stable operation even in sea states of 3 or higher.

Tested for real-world conditions

Before shipboard testing, the team verified performance using a full-scale simulation facility measuring 25 by 5 by 5 meters.

The facility replicated real ship compartments, including lighting and color conditions. Researchers recreated various fire and non-fire situations, such as lighters, welding sparks, and electric heaters, to train the AI for accurate fire identification.

The system successfully handled both open-area and shielded oil fires, including those likely to occur on aircraft carriers.

During testing, it extinguished a 4.5-square-meter open fire and a shielded fire beneath a helicopter-sized structure. These results proved its ability to respond to complex fire conditions at sea.

Real-ship tests were later conducted aboard the ROKS Ilchulbong, an LST-II class amphibious assault ship.

There, the system accurately targeted an oil fire 18 meters away, even in one-meter-high waves.

To maintain precision, KIMM developed a reinforcement learning algorithm that continuously adjusts the nozzle’s aiming angle using six degrees of freedom acceleration data to compensate for wave and hull movement.

Expanding safety beyond naval use

“This newly developed initial suppression firefighting system for shipboard oil fires is the world’s first technology to complete step-by-step verification from land-based simulation facilities to actual shipboard environments,” said Senior Researcher Hyuk Lee of KIMM.

“It can autonomously respond to the most dangerous oil fires on ships in both open and shielded conditions, marking a groundbreaking turning point for crew safety and preserving the ship’s combat effectiveness.”

He added that the system’s applications extend well beyond naval use.

“This technology is applicable not only to various naval vessels but also to ammunition depots, military supply warehouses, aircraft hangars, and offshore plants,” he said.

“Its future expansion to civilian ships and petrochemical facilities will significantly enhance fire safety at sea and in industrial settings.”

With its combination of AI precision, adaptive learning, and successful real-world testing, the KIMM system represents a significant advancement toward autonomous firefighting technologies for maritime and industrial safety.

🔗 Sumber: interestingengineering.com