In continuous kraft pulping, key quality indicators such as Kappa number and pulp viscosity are usually measured offline at low frequency, which limits real-time quality monitoring and control. Although data-driven soft sensors have shown potential for quality prediction, their performance is often restricted by limited labeled data and weak physical consistency. In addition, existing studies have focused mainly on single-target prediction, while the coupled prediction of Kappa number and pulp viscosity remains insufficiently explored despite their common dependence on cooking conditions and degradation kinetics. To address these issues, this study proposes a physics-guided multi-task learning framework (PG-MTL) for simultaneous prediction of Kappa number and pulp viscosity. The model combines a hard-parameter-sharing multi-task architecture with a physics-guided monotonicity constraint that enforces the expected non-increasing Kappa trend with increasing H-factor. Homoscedastic uncertainty weighting is also used to balance the two regression tasks during optimization. Experiments on industrial operating data show that PG-MTL achieved $R^2$= 0.920 for the Kappa number and $R^2$ = 0.910 for pulp viscosity. Compared with the strongest benchmark model, RMSE was reduced by 23.2% and 29.5% for the Kappa number and pulp viscosity, respectively. These results demonstrate that PG-MTL provides an effective and physically consistent solution for pulp-quality soft sensing under small-sample industrial conditions.

Cherry blossom extract (CBE) is rich in natural antioxidants; however, the poor release of bound bioactive compounds and the instability of the extract itself limit its further application. This study aimed to improve the antioxidant activity and stability of CBE by combining mixed-culture solid-state fermentation with microencapsulation. Using total phenolic content (TPC) as the primary response variable, we optimized the solid-state fermentation conditions for a mixed culture of Bacillus subtilis and Monascus purpureus via single-factor experiments combined with the Box–Behnken response surface method. Extracts prepared under these optimal conditions were then microencapsulated using a gelatin–chitosan co-precipitation method, and their antioxidant activity and in vitro release behavior were evaluated. The optimal fermentation conditions were an inoculation amount of 10.5%, an inoculation proportion of 1.3:1 (w/w), and a fermentation time of 7.3 days, producing a total phenolic content (TPC) of 38.49 ± 0.41 mg gallic acid equivalent (GAE)/g dry weight (DW). After 150 h, the micro-capsules had released 49.6% of their bioactive contents, with 50.4% remaining encapsulated. Collectively, these findings demonstrate that the mixed-culture fermentation extract exhibited higher antioxidant activity than both the single-strain extract and the non-fermented control, thereby overcoming key application bottlenecks of CBE.