
ICDS co-hire talks DeepSeek AI
Posted on February 4, 2025Editor’s Note: A version of this article was published on Penn State News.
UNIVERSITY PARK, Pa. — Dana Calacci, assistant professor of information sciences and technology and Penn State Institute for Computational and Data Sciences (ICDS) co-hire, recently spoke to Penn State News about DeepSeek’s approach to generative artificial intelligence (AI).
DeepSeek, an AI assistant that is designed to accomplish a variety of tasks, is advertised to be strong at formal reasoning tasks such as math and logic problems, according to Shomir Wilson, associate professor of information sciences and technology.
“This is a significant claim because many contemporary AI assistants struggle with formal reasoning tasks,” Wilson told Penn State News.
The system recently made headlines for having a fraction of the cost and computing power as some of its competitors.
“I think the approach the DeepSeek team takes is good for AI development for a number of reasons,” Calacci said. “The first is that right now, many models are evaluated against a “global” idea of what a “good” response is to a given query or prompt. This creates biases and makes models less useful for marginalized groups and specific tasks.”
Calacci suggests that DeepSeek allows for more “fine-grained, task-specific” training that is more affordable and faster than current approaches.
“This means that their strategy could be used to make models that, for some prompts, are more accurate or more useful to specific communities. This is an under-reported, interesting benefit of their approach,” Calacci said.
Calacci also notes that DeepSeek can run on more energy-efficient devices, potentially making their large language models more accessible without relying on cloud computing infrastructure.
“DeepSeek’s approach uses half as much compute as GPT-4 to train, which is major improvement,” Calacci added. “The size of the final DeepSeek model also means probably over a 90% reduction in the energy cost of a query compared to GPT-4, which is huge. Right now, GPT-4 queries are run on big cloud server infrastructure.”
Calacci’s research focuses on crowdsourced AI audits and AI harms, data tools for workers, data rights as labor rights and commercial surveillance.
Calacci and Wilson spoke to Penn State News alongside other University AI experts Akhil Kumar, professor of supply chain and information systems, and Mehmet Canayaz, assistant professor of finance.
Share
Related Posts
- Featured Researcher: Nick Tusay
- Multi-institutional team to use AI to evaluate social, behavioral science claims
- NSF invests in cyberinfrastructure institute to harness cosmic data
- Center for Immersive Experiences set to debut, serving researchers and students
- Distant Suns, Distant Worlds
- CyberScience Seminar: Researcher to discuss how AI can help people avoid adverse drug interactions
- AI could offer warnings about serious side effects of drug-drug interactions
- Taking RTKI drugs during radiotherapy may not aid survival, worsens side effects
- Cost-effective cloud research computing options now available for researchers
- Costs of natural disasters are increasing at the high end
- Model helps choose wind farm locations, predicts output
- Virus may jump species through ‘rock-and-roll’ motion with receptors
- Researchers seek to revolutionize catalyst design with machine learning
- Resilient Resumes team places third in Nittany AI Challenge
- ‘AI in Action’: Machine learning may help scientists explore deep sleep
- Clickbait Secrets Exposed! Humans and AI team up to improve clickbait detection
- Focusing computational power for more accurate, efficient weather forecasts
- How many Earth-like planets are around sun-like stars?
- Professor receives NSF grant to model cell disorder in heart
- SMH! Brains trained on e-devices may struggle to understand scientific info
- Whole genome sequencing may help officials get a handle on disease outbreaks
- New tool could reduce security analysts’ workloads by automating data triage
- Careful analysis of volcano’s plumbing system may give tips on pending eruptions
- Reducing farm greenhouse gas emissions may plant the seed for a cooler planet
- Using artificial intelligence to detect discrimination
- Four ways scholars say we can cut the chances of nasty satellite data surprises
- Game theory shows why stigmatization may not make sense in modern society
- Older adults can serve communities as engines of everyday innovation
- Pig-Pen effect: Mixing skin oil and ozone can produce a personal pollution cloud
- Researchers find genes that could help create more resilient chickens
- Despite dire predictions, levels of social support remain steady in the U.S.
- For many, friends and family, not doctors, serve as a gateway to opioid misuse
- New algorithm may help people store more pictures, share videos faster
- Head named for Ken and Mary Alice Lindquist Department of Nuclear Engineering
- Scientific evidence boosts action for activists, decreases action for scientists
- People explore options, then selectively represent good options to make difficult decisions
- Map reveals that lynching extended far beyond the deep South
- Gravitational forces in protoplanetary disks push super-Earths close to stars
- Supercomputer cluster donation helps turn high school class into climate science research lab
- Believing machines can out-do people may fuel acceptance of self-driving cars
- People more likely to trust machines than humans with their private info
- IBM donates system to Penn State to advance AI research
- ICS Seed Grants to power projects that use AI, machine learning for common good
- Penn State Berks team advances to MVP Phase of Nittany AI Challenge
- Creepy computers or people partners? Working to make AI that enhances humanity
- Sky is clearing for using AI to probe weather variability
- ‘AI will see you now’: Panel to discuss the AI revolution in health and medicine
- Privacy law scholars must address potential for nasty satellite data surprises
- Researchers take aim at hackers trying to attack high-value AI models
- Girls, economically disadvantaged less likely to get parental urging to study computers
- Seed grants awarded to projects using Twitter data
- Researchers find features that shape mechanical force during protein synthesis
- A peek at living room decor suggests how decorations vary around the world
- Interactive websites may cause antismoking messages to backfire
- Changing how government assesses risk may ease fallout from extreme financial events
- Penn State’s Leadership in AI Research
- Symposium at U.S. Capitol seeks solutions to election security
- ICS co-sponsors Health, Environment Seed Grant Program
- Differences in genes’ geographic origin influence mitochondrial function