Search

Continual Named Entity Recognition (CNER) is a burgeoning area, which involves updating an existing model by incorporating new entity types sequentially. Nevertheless, continual learning approaches are often severely afflicted by catastrophic forgetting. This issue is intensified in CNER due to the consolidation of old entity types from previous steps into the non-entity type at each step, leading to what is known as the semantic shift problem of the non-entity type. In this paper, we introduce a pooled feature distillation loss that skillfully navigates the trade-off between retaining knowledge of old entity types and acquiring new ones, thereby more effectively mitigating the problem of catastrophic forgetting. Additionally, we develop a confidence-based pseudo-labeling for the non-entity type, i.e., predicting entity types using the old model to handle the semantic shift of the non-entity type. Following the pseudo-labeling process, we suggest an adaptive re-weighting type-balanced learning strategy to handle the issue of biased type distribution. We carried out comprehensive experiments on ten CNER settings using three different datasets. The results illustrate that our method significantly outperforms prior state-of-the-art approaches, registering an average improvement of 6.3% and 8.0% in Micro and Macro F1 scores, respectively.

Realizing sustainable development has become a global priority. This holds, in particular, for agriculture. Recently, the United Nations launched the Sustainable Development Goals (SDGs), and the Nineteenth National People's Congress has delivered a national strategy for sustainable development in China-realizing green development. The overall objective of Agriculture Green Development (AGD) is to coordinate "green" with "development" to realize the transformation of current agriculture with high resource consumption and high environmental costs into a green agriculture and countryside with high productivity, high resource use efficiency and low environmental impact. This is a formidable task, requiring joint efforts of government, farmers, industry, educators and researchers. The innovative concept for AGD will focus on reconstructing the whole crop-animal production and food production-consumption system, with the emphasis on high thresholds for environmental standards and food quality as well as enhanced human well-being. This paper addresses the significance, challenges, framework, pathways and potential solutions for realizing AGD in China, and highlights the potential changes that will lead to a more sustainable agriculture in the future. Proposals include interdisciplinary innovations, whole food chain improvement and regional solutions. The implementation of AGD in China will provide important implications for the countries in developmental transition, and contribute to global sustainable development.

Grass-legume mixtures characterized by high productivity, efficient nitrogen (N) use and strong weed suppression were proposed to increase sustainability of grassland production. Adding competitive forbs to grassclover mixtures can be expected to further enhance the high productivity of grass-clover mixtures. This study investigates whether and to what extent adding forbs to a grass-clover mixture results in yield gains, and estimates the contributions of forb species (identity effect) and interactions with the grass-clover mixture (diversity effect) to these gains. It further examines the effectiveness of weed suppression in forb-containing grass-clover mixtures. Fifteen pure and mixed stands, all comprising the grass Lolium perenne L. and the legume Trifolium pratense L., and one or three competitive forb species (Cichorium intybus L., Carum carvi L., Plantago lanceolata L.) were established in a two-year field experiment and fertilised at two levels of N application. The diversity effect between P. lanceolata and the L. perenne-T. pratense mixture and a strong identity effect of P. lanceolata jointly contributed to yield gains of 10–21% over the binary L. perenne-T. pratense reference mixture for a wide range of P. lanceolata proportions across years and N fertilisation levels. In contrast, comparably smaller diversity and/or identity effects of C. intybus and C. carvi resulted in maximal yield gains of 5% and 7%, respectively, occurring at a narrower proportional range, but also in yield losses at high forb proportion. Including up to 80% forb in the L. perenne-T. pratense mixture effectively maintained weed suppression, but the degree of weed suppression decreased with increasing forb proportion in the following year. Including selected forbs, especially P. lanceolata, in productive grass-clover mixtures is promising for further enhancing productivity, meanwhile effectively suppressing short-term weed growth in intensively managed grasslands.

Anaerobic co-digestion of agricultural by-products or wastes with complementarity characteristics is commonly used to enhance methane yield. This study firstly explores the possibility of co-digesting grass and forb species (white clover, chicory and plantain) differing in nutrient composition in enhancing methane yield. This was examined with two inocula (a cattle manure-based inoculum and a grass-based inoculum) in a batch assay. Results showed that co-digesting grass and forbs synergistically enhanced methane yield potential on average by 31 L kg−1 volatile solids (+11%) and reduced lag phase time by 0.8 day in the grass-based inoculum, but not in the cattle manure-based inoculum. Mixtures containing plantain showed more consistent synergistic effect than chicory. Synergistic effects were attributed to more balanced nutrient composition (especially C/N ratio) in grass-forb mixtures. We demonstrate that anaerobic co-digestion of grass and forbs is feasible for enhancing methane yield, which promotes the utilization of multi-species grasslands for bioenergy production.

Wild flower strips is a common agri-environmental scheme used by farmers and land managers in order to improve biodiversity of pollinators. However, managed grasslands may also provide flower resources for flower visiting insects in agricultural landscapes. Botanically diverse grasslands on arable farms may support a range of wild pollinators, enhancing pollination services of crops. Intensively managed leys, on the other hand, typically contain only a few high-yielding, competitively strong species. One of the aims of the Multiplant project (2014-2018) was to test perennial seed mixtures targeted for bio-energy, feed protein and biodiversity, in order to develop multi-functional seed mixtures for grasslands. In the current study, we specifically investigated if yield (biomass production) and floral resources for pollinators could be simultaneously optimized by varying botanical composition of mixtures and cutting frequency. We tested four different perennial seed mixtures (3-, 5-, 11- and 13-species mixtures) at three sites varying in surrounding environment using three cutting strategies (no cutting, two cuts per year, four cuts per year). We measured flower production during the season, composition of flower-visitors (in functional groups), and biomass production of all plant species in the seed mixtures. The 11- and 13-species mixtures, which were designed to enhance pollinators, produced similar or higher yield than the 3- and 5- species mixtures under certain cutting regimes. The 3- and 5- species mixtures had a high accumulated flower abundance due to excessive flowering of lucerne under the two-cut strategy and white clover under the four-cut strategy. However, the 11- and 13 species mixtures presented a higher diversity of flowers during the flowering season. Interestingly, accumulated flower abundance was not significantly reduced under the two-cut strategy compared to no cut. Pollinator profiles (visits by different functional groups of insects) were plan

Enzyme–metal-ion–phosphate nanoflowers are high-surface area materials which are known to show higher activity than the constituting protein. Although the synthesis of hybrid nanoflowers has been demonstrated with a variety of proteins and reaction conditions, only di-valent metal ions have been tested to date. We expand on previous findings by testing a range of metal ions of different valence in co-presence with lipase from Burkholderia cepacia: Ag(I), Fe(II), Cu(II), Au(III). All metal ions produced colour precipitates, although the type of metal caused different precipitate morphologies under comparable reaction conditions: from nanoflowers to forests of nano-plates and crystal-like precipitates. In contrast, the type of metal ion did not appear to significantly affect the product's specific enzyme activity, which remained greater than that of free lipase. This indicates that the type of metal ion and the macroscopic arrangement of the petals play a secondary role to that of the co-presence of the metal and phosphate ions in determining lipase nanoflower activity. The demonstrated ability to produce metal–phosphate-protein nanoflowers with a selection of different metals also opens the way to producing a wider range of functional, nanostructured, materials.

Plant species diversity may benefit natural grassland productivity, but its effect in managed grassland systems is not well understood. A four-year multispecies grassland experiment was conducted to investigate the effect of species diversity-legumes and non-leguminous forbs-on productivity, persistence and sward quality under cutting or grazing regimes and with or without slurry application. Three mixtures were established- 3-mix: grass, red and white clover, 10-mix: 3-mix plus birdsfoot trefoil and six non-leguminous forbs, and 12-mix: 10-mix plus lucerne and festulolium. Species diversity increased sward production and yield persistence under cutting regime. The 12-mix had the highest yield from the second year onwards and no statistically significant yield reduction over four years, while annual yields in the 3-mix and 10-mix decreased significantly with increasing grassland age. The higher yield in the 12-mix was mainly due to the inclusion of high-yielding lucerne. The 10-mix and 12-mix had lower proportions of unsown species than the 3-mix, the difference being dependent on grassland age. Generally, the 3-mix had higher concentrations of in-vitro organic matter digestibility (IVOMD), neutral detergent fiber (NDF) and crude protein (CP), and a lower concentration of ash than the 10-mix and 12-mix. Slurry application increased annual yield production by 10% and changed the botanical composition, increasing the proportion of grass and decreasing the proportion of legumes. Compared to cutting, grazing increased forage production by 9% per cut on average and lowered legume and forb proportions in the mixtures, but yields did not differ among the three mixtures. Overall, our results suggest that species diversity increases sward productivity and persistence only under an ungrazed cutting regime. We conclude that increasing species diversity by selecting appropriate species with compatible management is key to achieving both high yields and high persistence in manag

Perennial ryegrass-clover mixtures are widely used in intensively managed grasslands due to their high biomass production. Recent studies have shown that inclusion of competitive forbs such as plantain enhanced above-and below-ground biomass of ryegrass-red clover mixtures (Cong et al., 2016), while inclusion of forbs such as caraway in the grass-clover mixture largely maintained the herbage yield. Meanwhile, inclusion of forbs increases plant litter diversity, which could change root quality of the plant community. Hence, changes in quantity and quality of root litter in forb-containing grass-clover mixtures may potentially affect soil microbial community structure, biomass and associated ecosystem functions, but it is yet to be elucidated. We hypothesized that inclusion of plantain in the grass-clover mixture would enhance soil microbial biomas and functions through its high biomass production and thus belowground carbon (C) input.