- Professor Min-Soo Kim’s research team in the School of Computing develops RL-SPH, a reinforcement-learning-based method that independently generates feasible solutions without relying on an external optimization solver to enforce feasibility

- Achieves 100% feasibility across five representative ILP benchmarks, with an average 28.6× lower primal gap than fully feasible baselines

- Trains 14.7× faster than representative learning-based approaches, demonstrating potential for AI-based decision-making in logistics, manufacturing, semiconductor production, and workforce management

Researchers developed a battery-free hydrogel bioelectronic patch that wirelessly monitors periodontal inflammation and releases miRNA therapy in response to disease-related enzyme activity. The device links local inflammation sensing with immune modulation, offering a new approach for integrated monitoring and treatment of periodontitis.

Structures exposed to wind, waves or flowing water can be difficult to model and monitor, especially when reliable measurements are scarce. A research team has developed FSI-PINNs, a physics-informed neural network framework that links fluid pressure with structural motion. Using no more than 1% of a simulated dataset, the method predicted flow fields and beam displacement and identified unknown structural density, offering a possible new tool for engineering problems in data-scarce environments.

Erosion removes more than soil

A paper published in Carbon Research reports that water erosion can intensify soil carbon loss not only by transporting soil downslope, but also by changing how microorganisms use the carbon that remains. Working in black soils of northeastern China, the team examined 0–100 cm soil profiles from erosion-prone upper slopes and depositional lower slopes to determine how erosion reshapes soil organic carbon (SOC) decomposition across depth.

The authors found that the strongest biological response occurred in the 0–40 cm topsoil. Upper slopes had lower SOC content and lower colloidal organic carbon, especially in the 2–10 μm fraction, than lower slopes. At the same time, SOC degradation rates in the upper slope were 41%–48% higher during incubation, indicating that erosion was associated with faster microbial conversion of soil carbon to CO₂.

Microbes adapt to carbon scarcity

The work combined field sampling, laboratory incubation, extracellular enzyme assays, 16S rRNA sequencing, and quantification of carbon-cycling functional genes. This integrated design allowed the authors to connect changes in carbon pools with microbial carbon limitation, community composition, and decomposition potential along slope position and soil depth.

According to the results, erosion reduced carbon availability in upper-slope topsoil and increased signs of microbial carbon limitation. Specific activities of carbon-, nitrogen-, and phosphorus-acquiring enzymes per unit SOC were higher in the upper slope, with β-glucosidase activity showing a marked increase. Microbial communities also shifted toward taxa associated with decomposition of more resistant substrates, particularly Streptomyces, which was enriched in the upper slope and positively correlated with SOC decomposition.

Functional genes point to recalcitrant carbon use

The upper slope also showed greater abundance of several genes linked to organic carbon breakdown in the top 40 cm, including genes associated with degradation of plant-derived compounds. The authors interpret this pattern as evidence that microorganisms under erosion stress reallocated effort toward accessing recalcitrant carbon substrates, rather than relying mainly on more readily available carbon pools.

A dual-purpose route for polluted water

Antibiotic residues in water are difficult to manage because treatment systems must remove persistent contaminants without adding major energy or chemical burdens. In Carbon Research, researchers led by Fan Li and Chun Hu describe a photocatalyst designed to address both issues at once: degrading emerging contaminants while producing H₂O₂ through oxygen reduction under visible light.

The material is a zinc single-atom catalyst anchored on graphitic carbon nitride, with asymmetrically coordinated Zn–N₂O₂ sites. According to the article, this coordination environment polarizes surface charge, broadens light absorption, and promotes directional electron transfer. The authors frame the approach as a way to couple water purification with value-added chemical generation, potentially lowering the overall energy demand of treatment.

Why asymmetric zinc sites matter

The team synthesized the catalyst by anchoring zinc single atoms onto graphitic carbon nitride and then characterized its structure and electronic properties using microscopy, X-ray photoelectron spectroscopy, extended X-ray absorption fine structure analysis, electrochemical measurements, and electron paramagnetic resonance. These results supported the presence of isolated zinc atoms coordinated by two nitrogen and two oxygen atoms.

That local structure was associated with stronger pollutant adsorption and improved charge behavior. Compared with undoped carbon nitride, the zinc-containing catalyst showed enhanced visible-light absorption, higher photocurrent, lower charge-transfer resistance, and reduced photoluminescence intensity, all consistent with more effective charge separation and transfer. The authors also report that ciprofloxacin adsorption on the modified catalyst was stronger than on pristine carbon nitride.