Compositional generalization is a basic mechanism in human language learning, but current neural networks lack such ability. In this paper, we conduct fundamental research for encoding compositionality in neural networks. Conventional methods use a single representation for the input sentence, making it hard to apply prior knowledge of compositionality. In contrast, our approach leverages such knowledge with two representations, one generating attention maps, and the other mapping attended input words to output symbols. We reduce the entropy in each representation to improve generalization. Our experiments demonstrate significant improvements over the conventional methods in five NLP tasks including instruction learning and machine translation. In the SCAN domain, it boosts accuracies from 14.0% to 98.8% in Jump task, and from 92.0% to 99.7% in TurnLeft task. It also beats human performance on a few-shot learning task. We hope the proposed approach can help ease future research towards human-level compositional language learning. Source code is available online 1 .

Zeptomole detector: A highly sensitive giant-magnetoresistive chip and FeCo nanoparticles can be used to linearly detect 600-4500 copies of streptavidin. Under unoptimized conditions, this system also detects human IL-6 with a sensitivity 13-times higher than that of standard ELISA techniques.

No abstract

A novel giant magnetoresistive sensor and uniform high-magnetic-moment FeCo nanoparticles (12.8 nm)-based detecting platform with minimized detecting distance was developed for rapid biomolecule quantification from body fluids. Such a system demonstrates specific, accurate, and quick detection and quantification of interleukin-6, a low-abundance protein and a potential cancer biomarker, directly in 4 muL of unprocessed human sera. This platform is expected to facilitate the identification and validation of disease biomarkers. It may eventually lead to a low-cost personal medical device for chronic disease early detection, diagnosis, and prognosis.

No abstract

Background: Artemisinin is a sesquiterpene lactone compound with a special peroxide bridge that is tightly linked to the cytotoxicity involved in fighting malaria and cancer. Artemisinin and its derivatives (ARTs) are considered to be potential anticancer drugs that promote cancer cell apoptosis, induce cell cycle arrest and autophagy, inhibit cancer cell invasion and migration. Additionally, ARTs significantly increase intracellular reactive oxygen species (ROS) in cancer cells, which results in ferroptosis, a new form of cell death, depending on the ferritin concentration. Ferroptosis is regarded as a cancer suppressor and is considered a new mechanism for cancer therapy. Methods: The anticancer activities of ARTs and reference molecules were compared by literature search and analysis. The latest research progress on ferroptosis was described, with a special focus on the molecular mechanism of artemisinin-induced ferroptosis. Results: Artemisinin derivatives, artemisinin-derived dimers, hybrids and artemisinin-transferrin conjugates, could significantly improve anticancer activity, and their IC50 values are lower than those of reference molecules such as doxorubicin and paclitaxel. The biological activities of linkers in dimers and hybrids are an important consideration in the drug design processes. ARTs induce ferroptosis mainly by triggering intracellular ROS production, promoting the lysosomal degradation of ferritin and regulating the System Xc-/Gpx4 axis. Interestingly, ARTs also stimulate the feedback inhibition pathway. Conclusion: Artemisinin and its derivatives could be used in the future as cancer therapies with broader application due to their induction of ferroptosis. Meanwhile, more attention should be paid to the development of novel artemisinin-related drugs based on the mechanism of artemisinin-induced ferroptosis.

A detection scheme for real-time Brownian relaxation of magnetic nanoparticles ͑MNPs͒ is demonstrated by a mixing-frequency method in this paper. MNPs are driven into the saturation region by a low frequency sinusoidal magnetic field. A high frequency sinusoidal magnetic field is then applied to generate mixing-frequency signals that are highly specific to the magnetization of MNPs. These highly sensitive mixing-frequency signals from MNPs are picked up by a pair of balanced built-in detection coils. The phase delays of the mixing-frequency signals behind the applied field are derived, and are experimentally verified. Commercial iron oxide MNPs with the core diameter of 35 nm are used for the measurement of Brownian relaxation. The results are fitted well with Debye model. Then a real-time measurement of the binding process between protein G and its antibody is demonstrated using MNPs as labels. This study provides a volume-based magnetic sensing scheme for the detection of binding kinetics and interaction affinities between biomolecules in real time.

No abstract