Issue link: https://iconnect007.uberflip.com/i/1533339
MARCH 2025 I PCB007 MAGAZINE 91 newsletter and Global SMT & Packaging's monthly "Business Outlook" column. He was a frequent presenter at IPC APEX EXPO over the years, as well as at EIPC, HKPCA, and other regional PCB events. His presentations were always informative, entertaining, and well attended. Walt said he became so active in IPC because of the continuous learning and networking opportunities. "IPC activities were always vibrant, challenging, and enjoyable," he told me. "ey offered constant learning experi- ences and interactions with interesting people. IPC's member base was diverse—culturally, geographically, and business-wise. I was able to make lasting friends and business contacts across the global supply chain." He became involved in roles that, he said, enabled him to "learn and contribute to both my employer and the global electronics industry." An early and lasting business goal for Walt was to understand the driving forces and cyclical nature of his employer's business. "is effort supported Morton Electronic Materials' sales and marketing, product development, financial forecasting, and pro- duction and inventory control activities," he said. "Key to this effort was collecting mar- ket data and developing useful forecasting methods with coincident and leading indica- tors for the various electronics-related mar- kets. Included in the data set were world and regional economic information with annual and monthly orders, and sales for electronic materials and components." When I was president of Dynachem, Walt was a VP on my staff, and I can attest to the accuracy and value of his work. His efforts have had an extensive impact on the indus- try. Walt's presentation and teaching of his research and information and the impact it had on IPC members and the industry at large, in addition to his decades of committee and group involvement within IPC, are key reasons Walt is an honored member of the IPC Hall of Fame. PCB007 Artificial intelligence (AI) is everywhere—from the chatbots we consult for customer support to tools predicting how diseases might spread. But the computing power and energy required to power modern AI models—such as large language mod- els (LLMs)—can make them expensive, inacces- sible, and environmentally taxing. A team of re- searchers at Rice University is working on solu- tions to change that. "Generative artificial intelligence is still in its in- fancy when it comes to broader integration," said Anshumali Shrivastava, associate professor of computer science, electrical and computer en- gineering and statistics. His research group pre- sented three of their most recent advancements in tweaking LLMs at the NeurIPS conference in Van- couver, British Columbia, last December. One paper explores "parameter sharing," in- troducing Sketch Structured Transforms (SS1)—a method for handling the vast tables of numbers, called weight matrices or working memory, that AI models rely on to make predictions and decisions. Shrivastava's team has developed an algorithm that allows LLMs to run efficiently on standard computer processors (CPUs) instead of GPUs. This work, outlined in a second paper leverages CPUs' very own hardware capabilities to redesign how calculations happen: The NoMAD Attention algo- rithm replaces complex operations with a clever al- ternative, using a feature of CPUs' memory archi- tecture in a way that's faster and less resource-in- tensive. This breakthrough means that in the near future, advanced AI tools might not just live in the cloud but could run directly on a phone or laptop. Shrivastava's work reflects a broader vision for the future of AI, one where advanced AI is avail- able to everyone, not just tech giants. (Source: Rice University) Rice Computer Scientists Develop Solutions for Making AI Models More Efficient and Customizable