Discovery Lectures

Tuesday - Discovery Lecture Recap

“Optimal Retail Tariff Design with Prosumers: Pursuing Equity at the Expenses of Economic Efficiencies”

Professor Yihsu Chen

On Tuesday, Professor Yihsu Chen introduced us to the economics behind the modern electricity grid and the challenges of designing fair electricity prices as more households begin generating their own power. He first explained how electricity is delivered from power plants through transmission and distribution systems before reaching consumers. Unlike traditional power systems, today's deregulated electricity markets separate electricity suppliers from utility companies, encouraging competition while allowing different organizations to generate, transmit, and distribute energy.

Professor Chen then introduced the idea of a prosumer: someone who both consumes and produces electricity (for example, as people generate electricity through their own rooftop solar panels on their homes). He explained how solar energy systems use inverters to convert electricity into a usable form and how any excess electricity can be sent back to the electrical grid. Depending on the pricing system, homeowners may receive credit for this excess energy through programs such as net metering or net billing, allowing them to offset some of their electricity costs.

The focus of the lecture was how these growing numbers of prosumers create new challenges for utility companies. Although households with solar panels purchase less electricity from the grid, utilities must still maintain power lines, transmission infrastructure, and other fixed costs that benefit all customers. If these costs are spread across fewer paying customers (as there are more prosumers), electricity prices can continue rising, encouraging even more people to install solar panels. This phenomenon is known as the utility death spiral. Professor Chen discussed his research on designing residential electricity tariffs that balance economic efficiency with energy equity, exploring combinations of usage-based charges and income-based fixed fees that ensure utilities can recover their costs while keeping electricity affordable for households with different income levels.

At the end of the lecture, Professor Chen showed how various mathematical optimization models can be used to evaluate different pricing strategies and identify policies that best balance fairness, affordability, and the long-term sustainability of the electrical grid. By the end of the lecture, we gained a better understanding of how economics, engineering, and public policy come together to shape the future of clean energy.

Written by Clarina Manuel

Wednesday - Discovery Lecture Recap

“Particle in a box in the carbon flatland”

Professor Jairo Velasco Jr.

Even on a hump day, COMOS students were tuned into Physics professor Jairo Velasco Jr.’s extraordinary discovery lecture on one of the oldest physics problems in the world: understanding the electron. In his speech, Velasco explored the ‘particle in a box’ problem, and how his lab worked to probe how electrons act with chaotic mechanics, and how it relates to one of the fastest growing fields yet: quantum materials. 

A particle in a box is an extremely interesting physics question relating to chaotic dynamics. Chaotic dynamics is the study of the change of motion over time. In its classical form, physics students study pendulum movement. For example, if you attached a string to a ball, and held that sting in place. That continuous movement is chaotic dynamics. But– how does that relate to the particle in a box theory?

A particle in a box is a type of chaotic dynamic. If you strike a ball in a circular container and let it bounce off its walls, it would create a similar circle in the middle. This is regular dynamics. But what if that same particle was in an oval container? Would it create a pattern? That's exactly what Professor Velasco’s group sought to answer. 

To set up his experiments, Velasco used graphene as a substrate and scanning tunneling spectroscopy. His lab carved out a stadium shaped corral using doping, and then sent charged electrons into the set up. The group was surprised to find that instead of an absence of pattern like in classical dynamics, there was noticeable sculpted potential. His potential revealed two types of “quantum scars.” Quantum scars are the enhancement of the probability density in the vicinity of a short classical unstable periodic orbit. Specifically, the team found that they adhere to energy relations for relativistic systems as expected for graphene electronic states. 

The properties of quantum scaring are profound. They can be used to steer electrons along desired paths, retain memory despite chaos, and localize wavefunctions in chaotic systems. All of which can be used from applications in mesoscopic physics and nanoelectronics, to quantum information and computing, and finally to quantum thermalization and ergodicity. While there is still a lot to learn about how electrons behave, Professor Velasco sparked a light in COSMOS students to research even the smallest corners of our world.  

Written by Emily Halmer-Lee

Thursday - Discovery Lecture Recap

“Brain-Inspired Artificial Intelligence”

Professor Jason Eshraghian

On Thursday, Professor Jason Eshraghian introduced us to the field of neuromorphic computing, which explores how the human brain can inspire the next generation of artificial intelligence. He began by comparing the energy consumption of modern AI hardware to the brain. While a GPU consumes around 100 watts of power when running AI models, the human brain performs many complex tasks using only about 10 watts (about the same amount of energy as a small light bulb). This enormous efficiency gap has motivated researchers to rethink how AI systems are designed.

Professor Jason explained that neuromorphic computing is not about copying the brain exactly, but instead borrowing useful ideas from biology. He pointed out that the brain is not always the perfect model for engineering (the same way airplanes were inspired by birds, without literally flapping their wings). We also build computers from completely different materials than biological neurons, and scientists still have much to learn about how the brain itself works. Instead, researchers identify the brain’s most efficient strategies and adapt them to solve computational problems.

One example of this is the event camera, a special type of camera that only records changes in a scene instead of capturing every single pixel in every frame like a traditional camera. Because it ignores information that remains unchanged, it produces far less data while responding much more quickly to motion. Professor Jason showed how event cameras are useful for applications such as autonomous vehicles and astronomy, where detecting movement is often more important than recording static images. He showed us video examples of event-based telescopes that captured sprites, which are brief lightning strikes high above thunderstorms, which are difficult to observe using conventional cameras.

Then, Professor Jason talked about how the brain processes information. Unlike artificial neural networks, where every neuron is active during every computation, biological neurons communicate using spikes. Spikes are brief electrical signals that are only generated when necessary, like when thinking about a specific topic/situation. Professor Jason explained that spiking neural networks mimic this behavior, allowing computers to process information while using much less energy. His research group developed snnTorch, an open-source Python library that makes it easier for researchers around the world to build and study spiking neural networks.

Professor Jason also showed us many applications of neuromorphic computing. One project uses spiking neural networks to analyze brain activity in patients with Parkinson’s disease and control neurostimulators that help reduce symptoms while minimizing energy consumption. Rather than continuously stimulating the brain, the system learns when stimulation is actually needed, making treatment both more effective and more energy efficient. He also introduced SpikeGPT, one of the first large language models built using spiking neural networks, which performs language generation using ~30 times fewer operations than similarly sized transformer models. He also shared several of his lab’s projects, including MatMul-Free Language Models, Ouro, and NeuroDump, all of which aim to make AI faster, more efficient, and capable of reasoning with significantly less computational power.

Professor Jason showed us that the future of artificial intelligence may not come from building larger and larger models, but from designing systems that learn from the efficiency of the human brain.

Written by Clarina Manuel

Friday - Discovery Lecture Recap

“Outsmarting Superbugs: New Tools for an Age of Antibiotic Resistance”

Professor Mauricio Rojas-Andrade

Friday was an exciting day for COSMOS students. They dove into the deep world of biochemistry, and a major growing problem– a new antibiotic resistance in viruses, an issue with the potential to harm all humans. 

Mauriciio Rojas-Andrade, a professor in the department of chemistry and biochemistry at UCSC, started his lecture with a surprising statistic. From 2025 to 2050, it is estimated that there will be 39 million human deaths attributable to antibiotic resistance, making it the leading cause of preventable deaths just before cancer. But, where did this problem arise? 

As Professor Rojas-Andrade explained, antibiotics over-use. After penicillin was discovered, and proven to be a safe cure for human illnesses, it became over-prescribed. Doctors would essentially hand out bottles to patients, and farmers started to use it on their sick livestock. Unfortunately, this overt overuse of antibiotics allowed for them to pollute into the environment. The pollution led to more pathogens encountering them, and building their own resistances to antibiotics, eventually coming back in full force. 

There is little scientists can do to prevent the resistance that antibodies adapted to our drugs, but create new antibiotics, explained Rojas-Andrade. But, drug creation and testing is an extremely time intensive and costly process, one that many pharmaceutical companies would prefer not to spend. And even when new antibiotics come onto the market, hospitals rarely buy in bulk and only prescribe it to the most harsh cases due to the absorbent costs. 

While a paradox, Professor Rojas-Andrade works to solve in his lab to synthesise more antibiotics to fight the growing antibodies of pathogens. Not only does he focus on synthesis, but also raising awareness, and encouraging people to eat animal products from un-antibiotic treated animals. 

Professor Rojas-Andrade showed COSMOS students the importance of the antibody paradox, and clearly illustrated what goes into the pharmaceutical process of producing antibiotics. Students walked away with a fuller understanding of a problem gripping medical teams today, and how they can use chemistry to prevent it. 

Written by Emily Halmer-Lee