Federated Data Quality Assessment Approach: Robust Learning With Mixed Label Noise.

Federated learning (FL) has been an effective way to train a machine learning model distributedly, holding local data without exchanging them. However, due to the inaccessibility of local data, FL with label noise would be more challenging. Most existing methods assume only open-set or closed-set noise and correspondingly propose filtering or correction solutions, ignoring that label noise can be mixed in real-world scenarios. In this article, we propose a novel FL method to discriminate the type of noise and make the FL mixed noise-robust, named FedMIN. FedMIN employs a composite framework that captures local-global differences in multiparticipant distributions to model generalized noise patterns. By determining adaptive thresholds for identifying mixed label noise in each client and assigning appropriate weights during model aggregation, FedMIN enhances the performance of the global model. Furthermore, FedMIN incorporates a loss alignment mechanism using local and global Gaussian mixture models (GMMs) to mitigate the risk of revealing samplewise loss. Extensive experiments are conducted on several public datasets, which include the simulated FL testbeds, i.e., CIFAR-10, CIFAR-100, and SVHN, and the real-world ones, i.e., Camelyon17 and multiorgan nuclei challenge (MoNuSAC). Compared to FL benchmarks, FedMIN improves model accuracy by up to 9.9% due to its superior noise estimation capabilities.

Electrostatic Fields Promote Methanogenesis More than Polarized Bioelectrodes in Anaerobic Reactors with Conductive Materials.

Direct interspecies electron transfer (DIET) is a breakthrough that can surpass the limitations of anaerobic digestion. Conductive materials and polarized bioelectrodes are known to induce DIET for methane production but are still challenging to apply at a field scale. Herein, compared to polarized bioelectrodes, electrostatic fields that promote DIET were investigated in an anaerobic reactor with conductive materials. As a conductive material, activated carbon enriched its surface with electroactive microorganisms to induce DIET (cDIET). cDIET improved the methane yield to 254.6 mL/g CODr, compared to the control. However, polarized bioelectrodes induced electrode-mediated DIET and biological DIET (bDIET), in addition to cDIET, improving the methane yield to 310.7 mL/g CODr. Electrostatic fields selectively promoted bDIET and cDIET for further methane production compared to the polarized bioelectrodes. As the contribution of DIET increased, the methane yield increased, and the substrate residue decreased, resulting in a significant improvement in methane production.

Full recycling of citric acid wastewater through anaerobic digestion, air-stripping and pH control.

Anaerobic digestion effluent (ADE) from the anaerobic digestion treatment of citric acid wastewater can be reused as a potential substitute for process water in the citric acid fermentation. However, excessive sodium contained in ADE significantly decreases citric acid production. In this paper, the inhibition mechanism of sodium on citric acid fermentation was investigated. We demonstrated that excessive sodium did not increase oxidative stress for Aspergillus niger, but reduced the pH of the medium significantly over the period 4-24 h, which led to lower activities of glucoamylase and isomaltase secreted by A. niger, with a decrease of available sugar concentration and citric acid production. ADE was pretreated by air-stripping prior to recycle and 18 g/L calcium carbonate was added at the start of fermentation to control the pH of the medium. The inhibition caused by ADE was completely alleviated and citric acid production substantially increased from 118.6 g/L to 141.4 g/L, comparable to the fermentation with deionized water (141.2 g/L). This novel process could decrease wastewater discharges and fresh water consumption in the citric acid industry, with benefit to the environment.

Nitrite and nitrate as electron acceptors for bioelectrochemical ammonium oxidation under electrostatic field.

Bioelectrochemical ammonium oxidation with nitrite and nitrate as electron acceptors was investigated in bulk solution exposed to electrostatic field. In a bioelectrochemical reactor, electroactive nitrogen removal bacteria including ammonium oxidizing exoelectrogens (AOE) and denitrifying electrotrophs (DNE) were enriched by electrostatic field of 0.2 V/cm in a bulk solution containing nitrite, nitrate, and ammonium. Ammonium was oxidized simultaneously with decreases in nitrite and nitrate as electron acceptors due to direct interspecies electron transfer between AOE and DNE. The specific ammonium oxidation rate was 48 mg NH4-N/g VSS.d when nitrate fraction was 1/3 in the electron acceptor composed of nitrite and nitrate. The specific ammonium oxidation rate gradually decreased with increasing nitrate fraction. However, it was still 24 mg NH4-N/g VSS.d when nitrate was the only electron acceptor. This indicates that nitrate can be used as an electron acceptor for bioelectrochemical ammonium oxidation, although it is a less effective than nitrite. This finding provides an advantage that strict nitritation which selectively produces nitrite from ammonium can be avoided when treating ammonia-rich wastewater in a bioelectrochemical reactor.