More than just rats and fleas added to the'perfect storm' plague. Photograph of the fresco Trionfo della Morte, taken at its original location in the Camposanto Monumentale in Pisa. The fresco, known as the "Triumph of Death" and attributed to the painter Buonamico Buffalmacco, is not precisely dated; scholarly estimates range from 1335 to 1350. While it does not depict the Black Death explicitly, the selected detail shows victims of an epidemic from diverse social backgrounds, their souls carried off by demons. Breakthroughs, discoveries, and DIY tips sent every weekday.

Federated Learning (FL) enables model development by leveraging data distributed across numerous edge devices without transferring local data to a central server. However, existing FL methods still face challenges when dealing with scarce and label-skewed data across devices, resulting in local model overfitting and drift, consequently hindering the performance of the global model. In response to these challenges, we propose a pioneering framework called FLea, incorporating the following key components: i) A global feature buffer that stores activation-target pairs shared from multiple clients to support local training. This design mitigates local model drift caused by the absence of certain classes; ii) A feature augmentation approach based on local and global activation mix-ups for local training. This strategy enlarges the training samples, thereby reducing the risk of local overfitting; iii) An obfuscation method to minimize the correlation between intermediate activations and the source data, enhancing the privacy of shared features. To verify the superiority of FLea, we conduct extensive experiments using a wide range of data modalities, simulating different levels of local data scarcity and label skew. The results demonstrate that FLea consistently outperforms state-of-the-art FL counterparts (among 13 of the experimented 18 settings, the improvement is over 5% while concurrently mitigating the privacy vulnerabilities associated with shared features. Code is available at https://github.com/XTxiatong/FLea.git.

Learning a global model by abstracting the knowledge, distributed across multiple clients, without aggregating the raw data is the primary goal of Federated Learning (FL). Typically, this works in rounds alternating between parallel local training at several clients, followed by model aggregation at a server. We found that existing FL methods under-perform when local datasets are small and present severe label skew as these lead to over-fitting and local model bias. This is a realistic setting in many real-world applications. To address the problem, we propose \textit{FLea}, a unified framework that tackles over-fitting and local bias by encouraging clients to exchange privacy-protected features to aid local training. The features refer to activations from an intermediate layer of the model, which are obfuscated before being shared with other clients to protect sensitive information in the data. \textit{FLea} leverages a novel way of combining local and shared features as augmentations to enhance local model learning. Our extensive experiments demonstrate that \textit{FLea} outperforms the start-of-the-art FL methods, sharing only model parameters, by up to $17.6\%$, and FL methods that share data augmentations by up to $6.3\%$, while reducing the privacy vulnerability associated with shared data augmentations.

Just a half-millimeter wide, the tiny crabs can bend, twist, crawl, walk, turn and even jump. The researchers also developed millimeter-sized robots resembling inchworms, crickets and beetles. Although the research is exploratory at this point, the researchers believe their technology might bring the field closer to realizing micro-sized robots that can perform practical tasks inside tightly confined spaces. The research will be published on Wednesday (May 25) in the journal Science Robotics. Last September, the same team introduced a winged microchip that was the smallest-ever human-made flying structure.

Fairness-aware learning aims at constructing classifiers that not only make accurate predictions, but do not discriminate against specific groups. It is a fast-growing area of machine learning with far-reaching societal impact. However, existing fair learning methods are vulnerable to accidental or malicious artifacts in the training data, which can cause them to unknowingly produce unfair classifiers. In this work we address the problem of fair learning from unreliable training data in the robust multisource setting, where the available training data comes from multiple sources, a fraction of which might be not representative of the true data distribution. We introduce FLEA, a filtering-based algorithm that allows the learning system to identify and suppress those data sources that would have a negative impact on fairness or accuracy if they were used for training. We show the effectiveness of our approach by a diverse range of experiments on multiple datasets. Additionally we prove formally that, given enough data, FLEA protects the learner against unreliable data as long as the fraction of affected data sources is less than half.

HarperDB was named after CEO Stephen Goldberg's dog, a five-year old adopted pup.HarperDB The so-called Internet of Things (IoT) is growing, exponentially, obviously. As an example of the machines that populate the IoT, modern aircraft are estimated to now fly with connected sensors monitoring as many as 5000 component elements per engine every second - and that's just the engines. For equipment engineers in aviation (and every other industry now digitally transforming) this means a lot of head scratching, some cool innovations and a lot of fine-grained physical tuning with a fair dose of engine grease. The same challenge also exists for information technologists supporting these systems. For software programmers and database engineers in every industry, making the IoT work means a lot of brain-aches, some super-cool innovations and a lot of fine-grained keyboard and screen based tuning, with a fair dose of'virtual' microprocessor engine grease (spoiler alert: microprocessors are built in clean room labs and rarely get oiled with lubricant).

When it comes to things that are ultrafast and lightweight, robots can't hold a candle to the fastest-jumping insects and other small-but-powerful creatures. New research published in the journal Science could help explain why nature still beats robots--and describes how machines might take the lead. The multi-institutional team of authors includes Associate Professor Sarah Bergbreiter, who studies microrobotics at the University of Maryland's A. James Clark School of Engineering. Take the smashing mantis shrimp, a small crustacean not much bigger than a thumb. Its hammer-like mouthparts can repeatedly deliver 69-mile-per-hour wallops more than 100 times faster than the blink of an eye to break open hard snail shells.

Mechanical power limitations emerge from the physical trade-off between force and velocity. Many biological systems incorporate power-enhancing mechanisms enabling extraordinary accelerations at small sizes. We establish how power enhancement emerges through the dynamic coupling of motors, springs, and latches and reveal how each displays its own force-velocity behavior. We mathematically demonstrate a tunable performance space for spring-actuated movement that is applicable to biological and synthetic systems. Incorporating nonideal spring behavior and parameterizing latch dynamics allows the identification of critical transitions in mass and trade-offs in spring scaling, both of which offer explanations for long-observed scaling patterns in biological systems.