Upcoming Events

HKAE TechTalk – Transforming Wastewater into Green Hydrogen: An Option to Produce the Fuel of the Future in HK?

Speaker:  Professor Irene Lo, Chair Professor, Department of Civil and Environmental Engineering, HKUST

Abstract:

Conventional wastewater treatment facilities face significant challenges such as high carbon emissions and energy consumption. Our research team developed a low-carbon-emission photoelectrochemical (PEC) system for the treatment of saline wastewater, coupled with green hydrogen (H₂) generation. The PEC system demonstrates exceptional performance in treating saline sewage, effectively meeting Hong Kong discharge standards for chemical oxygen demand (COD), ammonia-N, and E. coli within 40 minutes at 2 V (vs. Ag/AgCl) when treating simulated saline sewage. When applied to real saline sewage (sampled from Stonecutters Island Sewage Treatment Works in Hong Kong), which has a more complex composition and higher turbidity, the PEC system still meets the discharge standards within 2 hours of PEC treatment. Additionally, we demonstrate the scalability of the PEC system from batch to a large-scale continuous flow reactor, and it can generate 18.13 mol/m³ of green H₂ (equivalent to 0.92 kWh/m³ of electricity). This innovative technology results in significant reductions of 98% in scope 1 emissions (direct GHG emissions) and 97% in total carbon emissions compared to conventional wastewater treatment plants, illustrating its excellent environmental benefits and enormous practical potential.

Language: English
Mode: Mixed

All members of the HKU members and the general public are welcome to join. Seats for on-site participants are limited.

For more details: https://innowings.engg.hku.hk/greenhydrogen/

[TechTalk] Building Multi-dimensional Parallel Training Systems for Large AI Models

Speaker:  Professor Heming Cui, 
Associate Professor, School of Computing and Data Science, HKU

About the talk:

The increasing modeling capacities of large DNNs (e.g., Transformer and GPT) have achieved unprecedented successes in various AI areas, including understanding vision and natural languages. The high modeling power a large DNN mainly stems from its increasing complexity (having more neuron layers and more neuron operators in each layer) and dynamicity (frequently activating/deactivating neuron operators in each layer during training, such as Neural Architecture Search, or NAS). Dr. Cui’s talk will present his recent papers (e.g., [PipeMesh, in revision of a journal], [Fold3D TPDS 2023], [NASPipe ASPLOS 2022], and [vPipe TPDS 2021]), which address major limitations in existing multi-dimensional parallel training systems, including GPipe, Pipedream, and Megatron. Fold3D is now the major thousands-GPU parallel training system on the world-renowned MindSpore AI framework.

Language: English
Mode: Mixed

All members of the HKU members and the general public are welcome to join. Seats for on-site participants are limited.

For more details:  https://innowings.engg.hku.hk/aimodels/