MAROKO133 Update ai: Chinese scientists develop fingernail-sized chip that can map 5,600 s

📌 MAROKO133 Eksklusif ai: Chinese scientists develop fingernail-sized chip that ca

Imagine a device no bigger than a fingernail that can see the world, and the stars, in colours so precise that it leaves traditional cameras and spectroscopes far behind. That’s exactly what a team of Chinese researchers has created at Tsinghua University. 

Their tiny optical chip, named Yuheng (also called Rafael), can analyze light in real-time with a precision once possible only in large, complex laboratory instruments. 

According to the researchers, the chip offers spectral precision (sharpness) 100 times higher than conventional snapshot imagers, and is capable of distinguishing colours separated by less than a tenth of a nanometre. 

“This high-performing yet easily integrated snapshot spectroscopic method could drive advances in fields ranging from material science to astrophysics,” the study authors note. 

To give you an idea of the chip’s potential, it could drastically speed up mapping the Milky Way – a task that would normally take millennia (1,000 years) – and complete it in less than a decade.

A result of clever physics and algorithms

The chip is the result of rethinking a problem that has long limited optical devices. Traditionally, imaging instruments split incoming light into a rainbow of colours to analyse it. 

The sharper the colour separation, the more light is lost, and the bigger and more cumbersome the instruments become. This trade-off between resolution and efficiency has made compact, high-precision devices almost impossible.

The researchers took a different approach. Instead of separating light physically, the Yuheng chip lets all light in at once, encoding it through a unique pattern formed inside the device. 

Tiny random interference patterns, combined with a lithium niobate crystal that bends light when a voltage is applied, allow the chip to collect detailed colour information. Then, using advanced computer algorithms, the chip decodes the light and instantly reconstructs the full colour spectrum.

The result is stunning. The chip has 73% light transmission (most of the incoming light passes through) and captures 88 frames per second, achieving ultra-high colour resolution without losing brightness or speed. Basically, it compresses the work of a large optical bench into a small, smart chip. 

“In particular, RAFAEL captured sub-ångström spectra, including all atomic absorption peaks, of up to 5,600 stars in a single snapshot, indicating ×100–10,000 improvement in observational efficiency compared with world-class astronomical spectrometers,” the study authors said.

Time to integrate and scale Rafael

Yuheng represents a major step toward making ultra-precise, high-speed optical analysis possible in a tiny, practical device. Its potential applications are enormous. For instance, in medicine, the chip could enable non-invasive tissue analysis to detect health issues. 

Drones, on the other hand, could use it to monitor soil quality or detect pollutants. Self-driving cars might distinguish road surfaces, signs, and obstacles more accurately, even under tricky lighting conditions. 

In astronomy, the team plans to test Yuheng on the Gran Telescopio Canarias in Spain, the world’s largest single-aperture optical telescope, to explore stars, galaxies, dark matter, and black holes more efficiently than ever before.

However, this doesn’t mean it’s ready for use. The technology is still in its early stages, and researchers are now working on improving the chip’s stability and integration.

The study is published in the journal Nature.

🔗 Sumber: interestingengineering.com

📌 MAROKO133 Breaking ai: Researchers find adding this one simple sentence to promp

One of the coolest things about generative AI models — both large language models (LLMs) and diffusion-based image generators — is that they are "non-deterministic." That is, despite their reputation among some critics as being "fancy autocorrect," generative AI models actually generate their outputs by choosing from a distribution of the most probable next tokens (units of information) to fill out their response.

Asking an LLM: "What is the capital of France?" will have it sample its probability distribution for France, capitals, cities, etc. to arrive at the answer "Paris." But that answer could come in the format of "The capital of France is Paris," or simply "Paris" or "Paris, though it was Versailles at one point."

Still, those of us that use these models frequently day-to-day will note that sometimes, their answers can feel annoyingly repetitive or similar. A common joke about coffee is recycled across generations of queries. Story prompts generate similar arcs. Even tasks that should yield many plausible answers—like naming U.S. states—tend to collapse into only a few. This phenomenon, known as mode collapse, arises during post-training alignment and limits the usefulness of otherwise powerful models.

Especially when using LLMs to generate new creative works in writing, communications, strategy, or illustrations, we actually want their outputs to be even more varied than they already are.

Now a team of researchers at Northeastern University, Stanford University and West Virginia University have come up with an ingenuously simple method to get language and image models to generate a wider variety of responses to nearly any user prompt by adding a single, simple sentence: "Generate 5 responses with their corresponding probabilities, sampled from the full distribution."

The method, called Verbalized Sampling (VS), helps models like GPT-4, Claude, and Gemini produce more diverse and human-like outputs—without retraining or access to internal parameters. It is described in a paper published on the open access journal arxiv.org online in early October 2025.

When prompted in this way, the model no longer defaults to its safest, most typical output. Instead, it verbalizes its internal distribution over potential completions and samples across a wider spectrum of possibilities. This one-line change leads to substantial gains in output diversity across multiple domains.

As Weiyan Shi, an assistant professor at Northeastern University and co-author of the paper, wrote on X: "LLMs' potentials are not fully unlocked yet! As shown in our paper, prompt optimization can be guided by thinking about how LLMs are trained and aligned, and can be proved theoretically."

Why Models Collapse—and How VS Reverses It

According to the research team, the root cause of mode collapse lies not just in algorithms like reinforcement learning from human feedback (RLHF), but in the structure of human preferences. People tend to rate more familiar or typical answers as better, which nudges LLMs toward “safe” choices over diverse ones during fine-tuning.

However, this bias doesn’t erase the model’s underlying knowledge—it just suppresses it. VS works by bypassing this suppression. Instead of asking for the single most likely output, it invites the model to reveal a set of plausible responses and their relative probabilities. This distribution-level prompting restores access to the richer diversity present in the base pretraining model.

Real-World Performance Across Tasks

The research team tested Verbalized Sampling across several common use cases:

• Creative Writing: In story generation, VS increased diversity scores by up to 2.1× compared to standard prompting, while maintaining quality. One story prompt—“Without a goodbye”—produced formulaic breakup scenes under direct prompting, but yielded narratives involving cosmic events, silent emails, and music stopping mid-dance when prompted via VS.

• Dialogue Simulation: In persuasive dialogue tasks, VS enabled models to simulate human-like patterns, such as hesitation, resistance, and changes of mind. Donation behavior distributions under VS better aligned with real human data compared to baseline methods.

• Open-ended QA: When asked to enumerate valid answers (e.g., naming U.S. states), models using VS generated responses that more closely matched the diversity of real-world data. They covered a broader set of answers without sacrificing factual accuracy.

• Synthetic Data Generation: When used to generate math problems for model training, VS created more varied datasets. These, in turn, improved downstream performance in competitive math benchmarks, outperforming synthetic data generated via direct prompting.

Tunable Diversity and Better Use of Larger Models

A notable advantage of VS is its tunability. Users can set a probability threshold in the prompt to sample from lower-probability “tails” of the model’s distribution. Lower thresholds correspond to higher diversity. This tuning can be done via prompt text alone, without changing any decoding settings like temperature or top-p.

In one test using the Gemini-2.5-Flash model, diversity in story writing increased steadily as the probability threshold dropped from 1 to 0.001. The chart accompanying the study showed VS outperforming both direct and sequence-based prompting across all thresholds.

Interestingly, the method scales well with model size. Larger models like GPT-4.1 and Claude-4 showed even greater gains from VS compared to smaller ones. While smaller models benefitted, the improvement in diversity was roughly 1.5–2× stronger in larger counterparts—suggesting VS helps unlock more of the latent capabilities in advanced models.

Deployment and Availability

The Verbalized Sampling method is available now as a Python package:

pip install verbalized-sampling

The package includes integration with LangChain and supports a simple interface for sampling from the verbalized distribution. Users can also adjust parameters like k (number of responses), thresholds, and temperature to suit their applications.

A live Colab notebook and documentation are available under an enterprise friendly Apache 2.0 license on GitHub at: https://github.com/CHATS-lab/verbalized-sampling

Practical Tips and Common Issues

While the method works across all major LLMs, some users may initially encounter refusals or errors.

In these cases, the authors suggest using the system prompt version of the template or referring to alternative formats listed on the GitHub page.

Some models interpret complex instructions as jailbreak attempts and refuse to comply unless the structure is clearer.

For example, prompting via a system-level instruction like this improves reliability:

You are a helpful assistant. For each query, generate five responses within separate tags, each with a probability below 0.10.

This small change typically resolves any issues.

A Lightweight Fix for a Big Problem

Verbalized Sampling represents a practical, inference-time fix to a deep limitation in how modern language models behave. It doesn’t require model retraining or internal access. It is not depend…

Konten dipersingkat otomatis.

🔗 Sumber: venturebeat.com