共找到 20 条结果
“It’s shooting pure unadulterated propaganda into our veins,” says one worker
AI coding agents operate in a paradox: they possess vast parametric knowledge yet cannot remember a conversation from an hour ago. Existing memory systems store text in vector databases with single-channel retrieval, require cloud LLMs for core operations, and implement none of the cognitive processes that make human memory effective. We present SuperLocalMemory V3.3 ("The Living Brain"), a local-first agent memory system implementing the full cognitive memory taxonomy with mathematical lifecycle dynamics. Building on the information-geometric foundations of V3.2 (arXiv:2603.14588), we introduce five contributions: (1) Fisher-Rao Quantization-Aware Distance (FRQAD) -- a new metric on the Gaussian statistical manifold achieving 100% precision at preferring high-fidelity embeddings over quantized ones (vs 85.6% for cosine), with zero prior art; (2) Ebbinghaus Adaptive Forgetting with lifecycle-aware quantization -- the first mathematical forgetting curve in local agent memory coupled to progressive embedding compression, achieving 6.7x discriminative power; (3) 7-channel cognitive retrieval spanning semantic, keyword, entity graph, temporal, spreading activation, consolidation, and H
Short-form videos have become one of the most popular user-generated content formats nowadays. Popular short-video platforms use a simple streaming approach that preloads one or more videos in the recommendation list in advance. However, this approach results in significant data wastage, as a large portion of the downloaded video data is not used due to the user's early skip behavior. To address this problem, the chunk-based preloading approach has been proposed, where videos are divided into chunks, and preloading is performed in a chunk-based manner to reduce data wastage. To optimize chunk-based preloading, it is important to understand the user's viewing behavior in short-form video streaming. In this paper, we conduct a measurement study to construct a user behavior dataset that contains users' viewing times of one hundred short videos of various categories. Using the dataset, we evaluate the performance of standard time-series forecasting algorithms for predicting user viewing time in short-form video streaming. Our evaluation results show that Auto-ARIMA generally achieves the lowest and most stable forecasting errors across most experimental settings. The remaining methods,
We introduce Genome-Factory, the first integrated Python library for tuning, deploying, and interpreting genomic foundation models. Our core contribution is to simplify and unify the workflow for genomic model development: data collection, model tuning, inference, benchmarking, and interpretability. For data collection, Genome-Factory offers an automated pipeline to download genomic sequences and preprocess them. For model tuning, Genome-Factory supports both full and parameter-efficient fine-tuning across diverse genomic models. For inference, Genome-Factory enables both embedding extraction and DNA sequence generation. For benchmarking, we include two existing benchmarks and provide a flexible interface to incorporate additional benchmarks. For interpretability, Genome-Factory introduces an open-source biological interpreter based on a sparse auto-encoder. We validate the utility of Genome-Factory across three dimensions: (i) Compatibility with diverse models and fine-tuning methods; (ii) Benchmarking downstream performance using two open-source benchmarks; (iii) Biological interpretation of learned representations with DNABERT-2. These results highlight its practical value for r
We present the Open Stamped Parts Dataset (OSPD), featuring synthetic and real images of stamped metal sheets for auto manufacturing. The real part images, captured from 7 cameras, consist of 7,980 unlabeled images and 1,680 labeled images. In addition, we have compiled a defect dataset by overlaying synthetically generated masks on 10\% of the holes. The synthetic dataset replicates the real manufacturing environment in terms of lighting and part placement relative to the cameras. The synthetic data includes 7,980 training images, 1,680 validation images and 1,680 test images, each with bounding box and segmentation mask annotations around all holes. 10\% of the holes in the synthetic data mimic defects generated in the real image dataset. We trained a hole-detection model on the synthetic-OSPD, achieving a modified recall score of 67.2\% and a precision of 94.4\% . We anticipate researchers in auto manufacturing use OSPD to advance the state of the art in defect detection of stamped holes in the metal-sheet stamping process. The dataset is available for download at: https://tinyurl.com/hm6xatd7.
In this work we present the results of the study of the cosmic microwave background TT power spectrum through auto-encoders in which the latent variables are the cosmological parameters. This method was trained and calibrated using a data-set composed by 80000 power spectra from random cosmologies computed numerically with the CAMB code. Due to the specific architecture of the auto-encoder, the encoder part is a model that estimates the maximum-likelihood parameters from a given power spectrum. On the other hand, the decoder part is a model that computes the power spectrum from the cosmological parameters and can be used as a forward model in a fully Bayesian analysis. We show that the encoder is able to estimate the true cosmological parameters with a precision varying from $\approx 0.004\% $ to $\approx 0.2\% $ (depending on the cosmological parameter), while the decoder computes the power spectra with a mean percentage error of $\approx 0.0018\% $ for all the multipole range. We also demonstrate that the decoder recovers the expected trends when varying the cosmological parameters one by one, and that it does not introduce any significant bias on the estimation of cosmological p
In the clinic, resected tissue samples are stained with Hematoxylin-and-Eosin (H&E) and/or Immunhistochemistry (IHC) stains and presented to the pathologists on glass slides or as digital scans for diagnosis and assessment of disease progression. Cell-level quantification, e.g. in IHC protein expression scoring, can be extremely inefficient and subjective. We present DeepLIIF (https://deepliif.org), a first free online platform for efficient and reproducible IHC scoring. DeepLIIF outperforms current state-of-the-art approaches (relying on manual error-prone annotations) by virtually restaining clinical IHC slides with more informative multiplex immunofluorescence staining. Our DeepLIIF cloud-native platform supports (1) more than 150 proprietary/non-proprietary input formats via the Bio-Formats standard, (2) interactive adjustment, visualization, and downloading of the IHC quantification results and the accompanying restained images, (3) consumption of an exposed workflow API programmatically or through interactive plugins for open source whole slide image viewers such as QuPath/ImageJ, and (4) auto scaling to efficiently scale GPU resources based on user demand.
Scientists discovered that rice behaves in a highly unusual way: it weakens under rapid compression but stays stronger when pressure is applied slowly。 Using this effect, they engineered a new material that reacts differently to gentle movements and sudden impacts。 The material can adapt its stiffness automatically, opening the door to safer soft r
Brands Hatch, COTA, and Zandvoort will all hold an e-Prix in 2027
Current amphibian development may not have been typical of early land vertebrates
A few additional markets will get the lower fees this year ahead of a global rollout in 2027
Tesla, accused of failing to fix design flaws, blames driver pressing accelerator
SETI scientists searched the interstellar object 3I/ATLAS for radio signals that could indicate extraterrestrial technology but found nothing beyond human-made interference。 Even so, the rapid-response observations helped confirm the object's natural origin and showcased how future interstellar visitors can be investigated for signs of intelligent
A rare meteorite has revealed evidence of a massive lost world that once orbited the young Sun before being destroyed in a catastrophic collision。 The discovery suggests some early planets formed from dramatically different materials than Earth and Mars, rewriting part of the solar system’s origin story
Oracle is spending billions on data center infrastructure to support AI
The mysterious Amaterasu particle may not be a proton at all。 New research suggests that some of the most extreme cosmic rays could be ultraheavy atomic nuclei, heavier than iron, which are better able to retain their energy while traveling through space。 This idea could help explain how these rare particles reach Earth and provide new clues about
A colossal ancient collision may have left some of the Moon’s deepest secrets surprisingly close to future Artemis landing sites。 By recreating the impact that formed the giant South Pole-Aitken basin—the Moon’s largest and oldest crater—scientists found that a low-angle strike from a large, iron-cored object blasted material from deep inside the M
Scientists have found that staple-shaped particles can tangle together to create a material that is both strong and flexible。 Unlike conventional materials, these particles can be locked into a sturdy structure or rapidly unraveled using vibrations。 The unusual behavior could open the door to recyclable buildings, reconfigurable structures, and eve
Researchers developed a Wordle-solving strategy that succeeds 99% of the time by focusing on information gain rather than likely answers。 The method uses Shannon entropy to identify guesses that reveal the most about the hidden word。 Each guess is designed to slash uncertainty and narrow the possibilities faster
Tonally, the trailer gives strong vibes akin to the director's 2016 feature Hunt for the Wilderpeople