Persistent malnutrition is associated with poor clinical outcomes in cancer. However, assessing its reversibility can be challenging. The present study aimed to utilise machine learning (ML) to predict reversible malnutrition (RM) in patients with cancer. A multicentre cohort study including hospitalised oncology patients. Malnutrition was diagnosed using an international consensus. RM was defined as a positive diagnosis of malnutrition upon patient admission which turned negative one month later. Time-series data on body weight and skeletal muscle were modelled using a long short-term memory architecture to predict RM. The model was named as WAL-net, and its performance, explainability, clinical relevance and generalisability were evaluated. We investigated 4254 patients with cancer-associated malnutrition (discovery set = 2977, test set = 1277). There were 2783 men and 1471 women (median age = 61 years). RM was identified in 754 (17·7 %) patients. RM/non-RM groups showed distinct patterns of weight and muscle dynamics, and RM was negatively correlated to the progressive stages of cancer cachexia (r = –0·340, P < 0·001). WAL-net was the state-of-the-art model among all ML algorithms evaluated, demonstrating favourable performance to predict RM in the test set (AUC = 0·924, 95 % CI = 0·904, 0·944) and an external validation set (n 798, AUC = 0·909, 95 % CI = 0·876, 0·943). Model-predicted RM using baseline information was associated with lower future risks of underweight, sarcopenia, performance status decline and progression of malnutrition (all P < 0·05). This study presents an explainable deep learning model, the WAL-net, for early identification of RM in patients with cancer. These findings might help the management of cancer-associated malnutrition to optimise patient outcomes in multidisciplinary cancer care.

Time parallelization, also known as PinT (parallel-in-time), is a new research direction for the development of algorithms used for solving very large-scale evolution problems on highly parallel computing architectures. Despite the fact that interesting theoretical work on PinT appeared as early as 1964, it was not until 2004, when processor clock speeds reached their physical limit, that research in PinT took off. A distinctive characteristic of parallelization in time is that information flow only goes forward in time, meaning that time evolution processes seem necessarily to be sequential. Nevertheless, many algorithms have been developed for PinT computations over the past two decades, and they are often grouped into four basic classes according to how the techniques work and are used: shooting-type methods; waveform relaxation methods based on domain decomposition; multigrid methods in space–time; and direct time parallel methods. However, over the past few years, it has been recognized that highly successful PinT algorithms for parabolic problems struggle when applied to hyperbolic problems. We will therefore focus on this important aspect, first by providing a summary of the fundamental differences between parabolic and hyperbolic problems for time parallelization. We then group PinT algorithms into two basic groups. The first group contains four effective PinT techniques for hyperbolic problems: Schwarz waveform relaxation (SWR) with its relation to tent pitching; parallel integral deferred correction; ParaExp; and ParaDiag. While the methods in the first group also work well for parabolic problems, we then present PinT methods specifically designed for parabolic problems in the second group: Parareal; the parallel full approximation scheme in space–time (PFASST); multigrid reduction in time (MGRiT); and space–time multigrid (STMG). We complement our analysis with numerical illustrations using four time-dependent PDEs: the heat equation; the advection–diffusion equation; Burgers’ equation; and the second-order wave equation.

Machine learning has already shown promising potential in tiled-aperture coherent beam combining (CBC) to achieve versatile advanced applications. By sampling the spatially separated laser array before the combiner and detuning the optical path delays, deep learning techniques are incorporated into filled-aperture CBC to achieve single-step phase control. The neural network is trained with far-field diffractive patterns at the defocus plane to establish one-to-one phase-intensity mapping, and the phase prediction accuracy is significantly enhanced thanks to the strategies of sin-cos loss function and two-layer output of the phase vector that are adopted to resolve the phase discontinuity issue. The results indicate that the trained network can predict phases with improved accuracy, and phase-locking of nine-channel filled-aperture CBC has been numerically demonstrated in a single step with a residual phase of λ/70. To the best of our knowledge, this is the first time that machine learning has been made feasible in filled-aperture CBC laser systems.

Timing of food intake is an emerging aspect of nutrition; however, there is a lack of research accurately assessing food timing in the context of the circadian system. The study aimed to investigate the relation between food timing relative to clock time and endogenous circadian timing with adiposity and further explore sex differences in these associations among 151 young adults aged 18–25 years. Participants wore wrist actigraphy and documented sleep and food schedules in real time for 7 consecutive days. Circadian timing was determined by dim-light melatonin onset (DLMO). The duration between last eating occasion and DLMO (last EO-DLMO) was used to calculate the circadian timing of food intake. Adiposity was assessed using bioelectrical impedance analysis. Of the 151 participants, 133 were included in the statistical analysis finally. The results demonstrated that associations of adiposity with food timing relative to circadian timing rather than clock time among young adults living in real-world settings. Sex-stratified analyses revealed that associations between last EO-DLMO and adiposity were significant in females but not males. For females, each hour increase in last EO-DLMO was associated with higher BMI by 0·51 kg/m2 (P = 0·01), higher percent body fat by 1·05 % (P = 0·007), higher fat mass by 0·99 kg (P = 0·01) and higher visceral fat area by 4·75 cm2 (P = 0·02), whereas non-significant associations were present among males. The findings highlight the importance of considering the timing of food intake relative to endogenous circadian timing instead of only as clock time.

Predicting epidemic trends of coronavirus disease 2019 (COVID-19) remains a key public health concern globally today. However, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) reinfection rate in previous studies of the transmission dynamics model was mostly a fixed value. Therefore, we proposed a meta-Susceptible-Exposed-Infectious-Recovered-Susceptible (SEIRS) model by adding a time-varying SARS-CoV-2 reinfection rate to the transmission dynamics model to more accurately characterize the changes in the number of infected persons. The time-varying reinfection rate was estimated using random-effect multivariate meta-regression based on published literature reports of SARS-CoV-2 reinfection rates. The meta-SEIRS model was constructed to predict the epidemic trend of COVID-19 from February to December 2023 in Sichuan province. Finally, according to the online questionnaire survey, the SARS-CoV-2 infection rate at the end of December 2022 in Sichuan province was 82.45%. The time-varying effective reproduction number in Sichuan province had two peaks from July to December 2022, with a maximum peak value of about 15. The prediction results based on the meta-SEIRS model showed that the highest peak of the second wave of COVID-19 in Sichuan province would be in late May 2023. The number of new infections per day at the peak would be up to 2.6 million. We constructed a meta-SEIRS model to predict the epidemic trend of COVID-19 in Sichuan province, which was consistent with the trend of SARS-CoV-2 positivity in China. Therefore, a meta-SEIRS model parameterized based on evidence-based data can be more relevant to the actual situation and thus more accurately predict future trends in the number of infections.

To apply hydrotalcites more effectively to the problem of dye wastewater, the effects of divalent metal ions on the structure and stability of hydrotalcites, especially on their photocatalytic activity, were compared. In the present study, M/Cr hydrotalcites (M3Cr-CO3-LDHs) (in which M = Mg, Co, Ni, Cu, Zn), where the M/Cr molar ratio was 3, were prepared by the co-precipitation method. The structures and properties were characterized using powder X-ray diffraction (XRD), Fourier-transform infrared spectroscopy (FTIR), thermogravimetric-differential thermal analysis (TG-DTA), and UV-Visible diffuse reflectance spectroscopy (UV-Vis DRS). The results showed that five kinds of M3Cr-CO3-LDHs were synthesized successfully, and the layered structure of the samples obtained was regular and the crystal phase was single. When methylene blue (MB) solution was exposed to ZnCr-CO3-LDHs, H2O2, and visible light irradiation, more than 90.67% of the methylene blue (MB) was removed after 140 min. The photocatalytic activity of the samples was in the order: Co3Cr-CO3-LDHs > Mg3Cr-CO3-LDHs > Cu3Cr-CO3-LDHs > Zn3Cr-CO3-LDHs > Ni3Cr-CO3-LDHs. The results of a catalytic mechanism study showed that photocatalytic degradation of MB involved a demethylation reaction, with the reactive species containing •O2-, •OH, and h+.

The single ionization and dissociation of ethanol molecules induced by low-energy electrons (E0 = 90 eV) are investigated using multiparticle coincident momentum spectroscopy. By detecting two outgoing electrons (e1 and e2) and one fragment ion in coincidence, we obtain the energy deposition (E0 − E1 − E2) during electron ionization of the molecule, i.e., the binding energy spectra, for production of the different ionic fragments C2H5OH+, C2H4OH+, COH+, and H3O+. These data allow us to study the ionization channels for different ionic products. In particular, we focus on H3O+ as a product of double hydrogen migration. It is found that this channel mainly originates from the ionization of outer-valance orbitals (3a″,10a′, 2a″, 9a′, 8a′, 1a″, and 7a′). Additionally, there are minor contributions from the inner-valence orbitals such as 6a′, 5a′, and 4a′. Quantum chemistry calculations show two fragmentation pathways: concerted and sequential processes for formation of H3O+.

Energy loss of protons with 90 and 100 keV energies penetrating through a hydrogen plasma target has been measured, where the electron density of the plasma is about 1016 cm−3 and the electron temperature is about 1-2 eV. It is found that the energy loss of protons in the plasma is obviously larger than that in cold gas and the experimental results based on the Bethe model calculations can be demonstrated by the variation of effective charge of protons in the hydrogen plasma. The effective charge remains 1 for 100 keV protons, while the value for 90 keV protons decreases to be about 0.92. Moreover, two empirical formulae are employed to extract the effective charge.

A high-power all polarization-maintaining (PM) chirped pulse amplification (CPA) system operating in the 2.0 μm range is experimentally demonstrated. Large mode area (LMA) thulium-doped fiber (TDF) with a core/cladding diameter of 25/400 μm is employed to construct the main amplifier. Through dedicated coiling and cooling of the LMA-TDF to manage the loss of the higher order mode and thermal effect, a maximum average power of 314 W with a slope efficiency of 52% and polarization extinction ratio of 20 dB is realized. The pulse duration is compressed to 283 fs with a grating pair, corresponding to a calculated peak power of 10.8 MW, considering the compression efficiency of 88% and the estimated Strehl ratio of 89%. Moreover, through characterizing the noise properties of the laser, an integrated relative intensity noise of 0.11% at 100 Hz−1 MHz is obtained at the maximum output power, whereas the laser timing jitter is degraded by the final amplifier from 318 to 410 fs at an integration frequency of 5 kHz to 1 MHz, owing to the self-phase modulation effect-induced spectrum broadening. The root-mean-square of long-term power fluctuation is tested to be 0.6%, verifying the good stability of the laser operation. To the best of our knowledge, this is the highest average power of an ultrafast laser realized from an all-PM-fiber TDF-CPA system ever reported.

This paper provides nonparametric specification tests for the commonly used homogeneous and stable coefficients structures in panel data models. We first obtain the augmented residuals by estimating the model under the null hypothesis and then run auxiliary time series regressions of augmented residuals on covariates with time-varying coefficients (TVCs) via sieve methods. The test statistic is then constructed by averaging the squared fitted values, which are close to zero under the null and deviate from zero under the alternatives. We show that the test statistic, after being appropriately standardized, is asymptotically normal under the null and under a sequence of Pitman local alternatives. A bootstrap procedure is proposed to improve the finite sample performance of our test. In addition, we extend the procedure to test other structures, such as the homogeneity of TVCs or the stability of heterogeneous coefficients. The joint test is extended to panel models with two-way fixed effects. Monte Carlo simulations indicate that our tests perform reasonably well in finite samples. We apply the tests to re-examine the environmental Kuznets curve in the United States, and find that the model with homogenous TVCs is more appropriate for this application.

An all-fiber high-power linearly polarized chirped pulse amplification (CPA) system is experimentally demonstrated. Through stretching the pulse duration to a full width of approximately 2 ns with two cascaded chirped fiber Bragg gratings (CFBGs), a maximum average output power of 612 W is achieved from a high-gain Yb-doped fiber that has a core diameter of 20 μm with a slope efficiency of approximately 68% at the repetition rate of 80 MHz. At the maximum output power, the polarization degree is 92.5% and the M2 factor of the output beam quality is approximately 1.29; the slight performance degradations are attributed to the thermal effects in the main amplifier. By optimizing the B-integral of the amplifier and finely adjusting the higher-order dispersion of one of the CFBGs, the pulse width is compressed to 863 fs at the highest power with a compression efficiency of 72%, corresponding to a maximum compressed average power of 440.6 W, single pulse energy of 5.5 μJ and peak power of about 4.67 MW. To the best of our knowledge, this is the highest average power of a femtosecond laser directly generated from an all-fiber linearly polarized CPA system.

According to the positive time-discounting assumption of intertemporal decision-making, people prefer to undergo negative events in the future rather than in the present. However, negative discounting has been identified in the intertemporal choice and loss domains, which refers to people’s preference to experience negative events earlier rather than later. Studies have validated and supported the "anticipated dread" as an explanation for negative discounting. This study again explored the effect of anticipated dread on intertemporal choice using content analysis; that is, having participants identify anticipated dread among reasons for negative discounting. This study also validated the effect of anticipated dread on negative discounting by manipulating anticipated dread. This study adds empirical and direct evidence for the role of anticipated dread in negative discounting.

Intertemporal choices involve tradeoffs between outcomes that occur at different times. Most of the research has used pure gains tasks and the discount rates yielding from those tasks to explain and predict real-world behaviors and consequences. However, real decisions are often more complex and involve mixed outcomes (e.g., sooner-gain and later-loss or sooner-loss and later-gain). No study has used mixed gain-loss intertemporal tradeoff tasks to explain and predict real-world behaviors and consequences, and studies involving such tasks are also scarce. Considering that tasks involving a combination of gains and losses may yield different discount rates and that existing pure gains tasks do not explain or predict real-world outcomes well, this study conducted two experiments to compare the discount rates of mixed gain-loss intertemporal tradeoffs with those of pure gains or pure losses (Experiment 1) and to examine whether these tasks predicted different real-world behaviors and consequences (Experiment 2). Experiment 1 suggests that the discount rate ordering of the four tasks was, from highest to lowest, pure gains, sooner-loss and later-gain, pure losses, and sooner-gain and later-loss. Experiment 2 indicates that the evidence supporting the claim that the discount rates of the four tasks were related to different real-world behaviors and consequences was insufficient.

An 8-week feeding trial was conducted to investigate the effects of dietary vitamin D3 supplementation on the growth performance, tissue Ca and P concentrations, antioxidant capacity, immune response and lipid metabolism in Litopenaeus vannamei larvae. A total of 720 shrimp (initial weight 0·50 ± 0·01 g) were randomly distributed into six treatments, each of which had three duplicates of forty shrimp per duplicate. Six isonitrogenous and isolipidic diets were formulated to contain graded vitamin D3 (0·18, 0·23, 0·27, 0·48, 0·57 and 0·98 mg/kg of vitamin D3, measured) supplementation levels. The results revealed that L. vannamei fed diet containing 0·48 mg/kg of vitamin D3 achieved the best growth performance. Compared with the control group, supplementing 0·48 mg/kg of vitamin D3 significantly increased (P < 0·05) the activities of catalase, total antioxidative capacity, alkaline phosphatase and acid phosphatase in serum and hepatopancreas. Expression levels of antioxidant and immune-related genes were synchronously increased (P < 0·05). Carapace P and Ca concentrations were increased (P < 0·05) with the increased vitamin D3 supplementation levels. Further analysis of lipid metabolism-related genes expression showed that shrimp fed 0·48 mg of vitamin D3 per kg diet showed the highest value in the expression of lipid synthesis-related genes, while shrimp fed 0·98 mg of vitamin D3 per kg diet showed the highest value in the expression of lipolysis-related genes. In conclusion, the results of present study indicated that dietary supplementation of 0·48 mg/kg of vitamin D3 could increase Ca and P concentrations, improve antioxidant capacity and immune response, and influence lipid metabolism in L. vannamei.