From a particle smasher encircling the moon to an "impossible" laser, five scientists reveal the experiments they would run in a world powered purely by imagination In physics, breakthroughs are rare. Experiments are slow, expensive and often end up refining, rather than rewriting, our understanding of the universe. But what if the only constraint on scientific ambition were imagination? We asked five physicists to describe the kind of experiment they would do if they didn't have to worry about budgets, engineering limitations or political realities. Not because we expect any of it to happen soon - though in a few cases, momentum is building - but because it is revealing to see where their minds go when the usual boundaries are stripped away. One researcher wants to launch radio telescopes deep into space to probe dark matter with cosmic energy flashes. Others are dreaming of completely new kinds of particle accelerator or lasers that push the at bounds of the possible.

We investigate the short-context dominance hypothesis: that for most sequences, a small local prefix suffices to predict their next tokens. Using large language models as statistical oracles, we measure the minimum context length (MCL) needed to reproduce accurate full-context predictions across datasets with sequences of varying lengths. For sequences with 1-7k tokens from long-context documents, we consistently find that 75-80% require only the last 96 tokens at most. Given the dominance of short-context tokens, we then ask whether it is possible to detect challenging long-context sequences for which a short local prefix does not suffice for prediction. We introduce a practical proxy to MCL, called Distributionally Aware MCL (DaMCL), that does not require knowledge of the actual next-token and is compatible with sampling strategies beyond greedy decoding. Our experiments validate that simple thresholding of the metric defining DaMCL achieves high performance in detecting long vs. short context sequences. Finally, to counter the bias that short-context dominance induces in LLM output distributions, we develop an intuitive decoding algorithm that leverages our detector to identify and boost tokens that are long-range-relevant. Across Q&A tasks and model architectures, we confirm that mitigating the bias improves performance.

A startup's ads for controversial embryo tests hit the New York City subway. One day this fall, I watched an electronic sign outside the Broadway-Lafayette subway station in Manhattan switch seamlessly between an ad for makeup and one promoting the website Pickyourbaby.com, Inside the station, every surface was wrapped with more ads--babies on turnstiles, on staircases, on banners overhead. To his mind, one should be as accessible as the other. Nucleus is a young, attention-seeking genetic software company that says it can analyze genetic tests on IVF embryos to score them for 2,000 traits and disease risks, letting parents pick some and reject others. This is possible because of how our DNA shapes us, sometimes powerfully.

Recent works proposed amortizing the cost by learning generalized wave functions across different structures and compounds instead of solving each problem independently.

Uncertainty Quantification (UQ) is crucial for reliable image segmentation.

Large language models (LLMs) power modern AI applications, but processing sensitive data on untrusted servers raises privacy concerns. Homomorphic encryption (HE) enables computation on encrypted data for secure inference. However, neural text generation requires decoding methods like argmax and sampling, which are non-polynomial and thus computationally expensive under encryption, creating a significant performance bottleneck. We introduce cutmax, an HE-friendly argmax algorithm that reduces ciphertext operations compared to prior methods, enabling practical greedy decoding under encryption. We also propose the first HE-compatible nucleus (top-p) sampling method, leveraging cutmax for efficient stochastic decoding with provable privacy guarantees. Both techniques are polynomial, supporting efficient inference in privacy-preserving settings. Moreover, their differentiability facilitates gradient-based sequence-level optimization as a polynomial alternative to straight-through estimators. We further provide strong theoretical guarantees for cutmax, proving its convergence via exponential amplification of the gap ratio between the maximum and runner-up elements. Evaluations on realistic LLM outputs show latency reductions of 24x-35x over baselines, advancing secure text generation.

Interesting to see how well the proposed model would do under such zero-shot setup (i.e. GPT -2 accuracy is taken from their paper. The BERT paper reports that BooksCorpus and Wikipedia contain 0.8B and 2.5B words, respectively. For our processed data, BooksCorpus and Wikipedia contain 0.75B and 2B words, respectively. The implementation is the same as word embedding, i.e., a lookup "Segment 1", and "Segment 2") and feed it to model input, which indicates the segment of input tokens.

Deep learning is transforming microscopy, yet models often fail when applied to images from new instruments or acquisition settings. Conventional adversarial domain adaptation (ADDA) retrains entire networks, often disrupting learned semantic representations. Here, we overturn this paradigm by showing that adapting only the earliest convolutional layers, while freezing deeper layers, yields reliable transfer. Building on this principle, we introduce Subnetwork Image Translation ADDA with automatic depth selection (SIT-ADDA-Auto), a self-configuring framework that integrates shallow-layer adversarial alignment with predictive uncertainty to automatically select adaptation depth without target labels. We demonstrate robustness via multi-metric evaluation, blinded expert assessment, and uncertainty-depth ablations. Across exposure and illumination shifts, cross-instrument transfer, and multiple stains, SIT-ADDA improves reconstruction and downstream segmentation over full-encoder adaptation and non-adversarial baselines, with reduced drift of semantic features. Our results provide a design rule for label-free adaptation in microscopy and a recipe for field settings; the code is publicly available.