MAROKO133 Hot ai: China’s bizarre-looking submarine could work as surface vessel, house la

📌 MAROKO133 Update ai: China’s bizarre-looking submarine could work as surface ves

China has recently revealed an unusual naval vessel that blends characteristics of both a surface ship and a submarine, sparking considerable speculation among defense analysts.

Photographs from the Huangpu shipyard in Guangzhou show a dark, slender trimaran-style craft with a small sail-like structure and markings that resemble immersion or depth indicators, suggesting it may be capable of operating partially or fully below the surface, according to reports.

Because its superstructure is minimal and no clear missile tubes, hangars, or crew spaces are visible, experts believe it could be an unmanned or minimally crewed platform. The vessel’s purpose remains unclear, but several theories have emerged.

Semi-submersible arsenal ship

It might be a semi-submersible arsenal ship capable of launching missiles before slipping underwater, a drone-deployment platform for aerial or underwater systems, a covert transport for special-operations missions in littoral environments, or simply a testbed for new hybrid naval technologies.

The lack of official information adds to the mystery, but the craft’s appearance underscores China’s willingness to experiment with unconventional designs that could complicate detection and tracking by rival navies. As more imagery and operational details emerge, observers will be watching for hints about whether this is an isolated prototype or the first step toward a new class of stealthy, adaptive maritime platforms.

Unique propulsion system

The vessel’s submarine-like attributes are perhaps even more pronounced. These features may also include a propulsor at the rear, which would point to a pump-jet being fitted, a possible feature noted by Alex Luck, a journalist who closely follows the People’s Liberation Army Navy (PLAN). Pump-jets offer an array of advantages over traditional propellers, above all the ability to reach higher speeds without noisy cavitation — this means they can transit long distances around much more stealthily, reported The War Zone.

More obvious in this view is the submarine-like sail, which is fitted with a snorkel or possibly an antenna mast.

Weeks ago, the vessel was first widely observed in satellite imagery of the People’s Liberation Army Navy’s Huangpu Shipyard in Guangzhou. The hull appears to be around 65 metres in length. Its hull design is a trimaran — a slender central hull with outriggers — which is uncommon for traditional submarines or surface warships.

The vessel is painted in a dark grey or black color, reminiscent of submarine camouflage or low-visibility coatings. Its profile combines elements typical of both submarines and surface ships.

The superstructure is minimal and narrow, suggesting little space for a traditional crew, large radar/sensor arrays, or living quarters, according to the report.

Some analysts believe the vessel may use a pump-jet propulsor rather than a conventional open-propeller. Pump-jets are quieter (less cavitation), which would support stealthy movement, especially underwater or near-surface.

A leading theory is that this is the semi-submersible arsenal ship which has been speculated would be built. Rumours of this emerged on the Chinese internet in 2017. Although hung from published scientific research, much of the reporting at the time was highly speculative. Stories blended fan art with rehashed U.S. Navy ideas of arsenal ships, the concept of a vessel designed solely to carry a large quantity of land-attack missiles. Possibly China has finally built one, reported Naval News.

🔗 Sumber: interestingengineering.com

📌 MAROKO133 Hot ai: MIT Researchers Unveil “SEAL”: A New Step Towards Self-Improvi

The concept of AI self-improvement has been a hot topic in recent research circles, with a flurry of papers emerging and prominent figures like OpenAI CEO Sam Altman weighing in on the future of self-evolving intelligent systems. Now, a new paper from MIT, titled “Self-Adapting Language Models,” introduces SEAL (Self-Adapting LLMs), a novel framework that allows large language models (LLMs) to update their own weights. This development is seen as another significant step towards the realization of truly self-evolving AI.

The research paper, published yesterday, has already ignited considerable discussion, including on Hacker News. SEAL proposes a method where an LLM can generate its own training data through “self-editing” and subsequently update its weights based on new inputs. Crucially, this self-editing process is learned via reinforcement learning, with the reward mechanism tied to the updated model’s downstream performance.

The timing of this paper is particularly notable given the recent surge in interest surrounding AI self-evolution. Earlier this month, several other research efforts garnered attention, including Sakana AI and the University of British Columbia’s “Darwin-Gödel Machine (DGM),” CMU’s “Self-Rewarding Training (SRT),” Shanghai Jiao Tong University’s “MM-UPT” framework for continuous self-improvement in multimodal large models, and the “UI-Genie” self-improvement framework from The Chinese University of Hong Kong in collaboration with vivo.

Adding to the buzz, OpenAI CEO Sam Altman recently shared his vision of a future with self-improving AI and robots in his blog post, “The Gentle Singularity.” He posited that while the initial millions of humanoid robots would need traditional manufacturing, they would then be able to “operate the entire supply chain to build more robots, which can in turn build more chip fabrication facilities, data centers, and so on.” This was quickly followed by a tweet from @VraserX, claiming an OpenAI insider revealed the company was already running recursively self-improving AI internally, a claim that sparked widespread debate about its veracity.

Regardless of the specifics of internal OpenAI developments, the MIT paper on SEAL provides concrete evidence of AI’s progression towards self-evolution.

Understanding SEAL: Self-Adapting Language Models

The core idea behind SEAL is to enable language models to improve themselves when encountering new data by generating their own synthetic data and optimizing their parameters through self-editing. The model’s training objective is to directly generate these self-edits (SEs) using data provided within the model’s context.

The generation of these self-edits is learned through reinforcement learning. The model is rewarded when the generated self-edits, once applied, lead to improved performance on the target task. Therefore, SEAL can be conceptualized as an algorithm with two nested loops: an outer reinforcement learning (RL) loop that optimizes the generation of self-edits, and an inner update loop that uses the generated self-edits to update the model via gradient descent.

This method can be viewed as an instance of meta-learning, where the focus is on how to generate effective self-edits in a meta-learning fashion.

A General Framework

SEAL operates on a single task instance (C,τ), where C is context information relevant to the task, and τ defines the downstream evaluation for assessing the model’s adaptation. For example, in a knowledge integration task, C might be a passage to be integrated into the model’s internal knowledge, and τ a set of questions about that passage.

Given C, the model generates a self-edit SE, which then updates its parameters through supervised fine-tuning: θ′←SFT(θ,SE). Reinforcement learning is used to optimize this self-edit generation: the model performs an action (generates SE), receives a reward r based on LMθ′’s performance on τ, and updates its policy to maximize the expected reward.

The researchers found that traditional online policy methods like GRPO and PPO led to unstable training. They ultimately opted for ReST^EM, a simpler, filtering-based behavioral cloning approach from a DeepMind paper. This method can be viewed as an Expectation-Maximization (EM) process, where the E-step samples candidate outputs from the current model policy, and the M-step reinforces only those samples that yield a positive reward through supervised fine-tuning.

The paper also notes that while the current implementation uses a single model to generate and learn from self-edits, these roles could be separated in a “teacher-student” setup.

Instantiating SEAL in Specific Domains

The MIT team instantiated SEAL in two specific domains: knowledge integration and few-shot learning.

• Knowledge Integration: The goal here is to effectively integrate information from articles into the model’s weights.
• Few-Shot Learning: This involves the model adapting to new tasks with very few examples.

Experimental Results

The experimental results for both few-shot learning and knowledge integration demonstrate the effectiveness of the SEAL framework.

In few-shot learning, using a Llama-3.2-1B-Instruct model, SEAL significantly improved adaptation success rates, achieving 72.5% compared to 20% for models using basic self-edits without RL training, and 0% without adaptation. While still below “Oracle TTT” (an idealized baseline), this indicates substantial progress.

For knowledge integration, using a larger Qwen2.5-7B model to integrate new facts from SQuAD articles, SEAL consistently outperformed baseline methods. Training with synthetically generated data from the base Qwen-2.5-7B model already showed notable improvements, and subsequent reinforcement learning further boosted performance. The accuracy also showed rapid improvement over external RL iterations, often surpassing setups using GPT-4.1 generated data within just two iterations.

Qualitative examples from the paper illustrate how reinforcement learning leads to the generation of more detailed self-edits, resulting in improved performance.

While promising, the researchers also acknowledge some limitations of the SEAL framework, including aspects related to catastrophic forgetting, computational overhead, and context-dependent evaluation. These are discussed in detail in the original paper.

Original Paper: https://arxiv.org/pdf/2506.10943

Project Site: https://jyopari.github.io/posts/seal

Github Repo: https://github.com/Continual-Intelligence/SEAL

The post MIT Researchers Unveil “SEAL”: A New Step Towards Self-Improving AI first appeared on Synced.

🔗 Sumber: syncedreview.com