MAROKO133 Hot ai: Germany tests solar wastewater plant cutting pollution 90% for disaster

📌 MAROKO133 Update ai: Germany tests solar wastewater plant cutting pollution 90%

A research team in central Germany is testing a compact wastewater treatment plant designed to operate entirely on solar power.

Built for disaster and emergency settings, the system relies on a reactor, floating foam cubes, and microorganisms to clean polluted water.

Led by Prof. Dr.-Ing. Markus Röhricht at the Technical University of Central Hesse (THM), the project has moved from laboratory research to real-world trials.

Since September, the pilot plant has been running at a wastewater treatment facility in the town of Lollar.

The project, called EnsAK, has received €242,500 in funding from the Federal Ministry for Research, Technology, and Space.

It will run for two years. Industry partners Saygin & Stein and EMW filtertechnik GmbH are involved, alongside the Lollar-Staufenberg Zweckverband, which operates local wastewater infrastructure.

How the reactor works

The small treatment plant was installed in Lollar within a few days by Saygin & Stein.

Engineers designed it to function under crisis conditions, where infrastructure is damaged or power supplies are limited.

THM student Louis Müller manages the plant on-site. He visits every two days and more often if problems arise.

Müller studies for a master’s degree in Climate Protection, Environmental and Safety Engineering.

“The pre-treated wastewater from the treatment plant flows into the reactor. It has already been cleaned of coarse dirt in the screening building and passed through a grit and grease trap,” Müller explains.

Inside the reactor, foam cubes move continuously through the wastewater.

Microorganisms grow on the cube surfaces and form a regenerating biofilm.

As the biofilm degrades, sludge settles in a secondary clarifier.

After treatment, the water can be discharged into a river. In Lollar, it undergoes one additional treatment step.

Performance and testing

According to Röhricht, the pilot plant already meets strict German standards.

“With our experimental wastewater treatment plant, we can already comply with the strict limit for organic pollutant load that applies to large wastewater treatment plants in Germany.”

The system reduces chemical oxygen demand by 90 percent.

It also removes 60 to 70 percent of nitrogen, a key contributor to waterway over-fertilization.

Researchers test the treated water regularly.

Teams collect samples on Sundays, Tuesdays, and Thursdays. Each sample represents a full twelve-hour operating cycle.

The laboratory then analyzes the composite samples at THM.

Improving efficiency further

Parallel laboratory experiments aim to improve efficiency.

The team wants to reduce wastewater residence time from 16 hours to about ten hours.

Researchers also hope to cut down the number of foam cubes.

The cubes currently fill about 30 percent of the reactor volume.

Biotechnology student Nicolas Jost is studying alternative cube materials.

His work compares fine-pored and coarse-pored structures. He is also testing higher wastewater loads.

Jost will base his master’s thesis on these experiments.

The pilot plant will operate for a full year. Researchers want to measure performance across seasonal temperature changes.

Once optimized, the system could deploy quickly in regions affected by wars, natural disasters, or humanitarian crises, where clean water access becomes critical.

🔗 Sumber: interestingengineering.com

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

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