Highlights from the International Conference on AI and Statistics

The 26th International Conference on Artificial Intelligence and Statistics (AISTATS) gathered experts in computer science, AI, machine learning, and statistics. It took place in Valencia, Spain, from April 25th to April 27th. This event was a key venue for sharing knowledge and exploring new advancements.

Activities at the conference included invited talks, oral presentations, posters, and more. Topics covered ranged from deep learning to game theory. The focus on working across disciplines made AISTATS a hub for learning and innovation.

Criteo, the platinum sponsor, sent three researchers to present their papers. This highlighted their commitment to the growth of AI and statistics. It also added diverse views and insights to the discussions at AISTATS.

The conference was a crucial platform for displaying innovative work and networking. Next year’s event is set to bring even more breakthroughs and fascinating findings. It’s a significant occasion for the field’s development and progress.

Researchers’ Takeaways from AISTATS 2023 at Criteo

The Criteo team went to AISTATS 2023, a big event on Artificial Intelligence and Statistics. This conference was a place for researchers to share big ideas and learn about new things in their field. Here’s what the Criteo researchers found important:

Morgane Goibert: Origins of Low-Dimensional Adversarial Perturbations

Morgane Goibert talked about how AI can be tricked by small changes in data. She looked into making AI stronger against these tricks. She also looked at other studies about making AI safe from attacks.

Song Duong: Large Language Models for Zero-shot and Few-shot Classification

Song Duong shared his work on using big language models to understand and work with data in new ways. He showed how these models can guess right even with little information. He found exciting ideas on fighting cyber threats in machine learning too.

Imad Aouali: Bandits, Optimal Transport, and Trustworthy Machine Learning

Imad Aouali, a Ph.D. student, focused on decision-making, moving things smartly, and making AI reliable. He mentioned studies about making better choices when things are uncertain. He also talked about how to measure results without tracking users.

These were just a few highlights from AISTATS 2023. The Criteo team and others had deep talks and shared ideas. Their findings show the cool and important work happening at the crossroads of AI and statistics, leading to new discoveries.

MBZUAI’s Contributions at ICCV 2023

The International Conference on Computer Vision (ICCV) was hosted in Paris, France. It was an outstanding event for the computer vision community. This year’s conference saw a 34% rise in paper submissions – a record 8,260 in total. Of these, 2,161 papers were accepted for presentation.

Among the accepted submissions, 152 papers were selected for oral presentations. This provided a great platform for researchers. They got to highlight their work, share insights, and have engaging discussions.

MBZUAI stood out at ICCV 2023 by presenting an incredible 30 papers. This achievement reflects the institution’s commitment to computer vision. It also highlights the skill and creativity of its researchers.

Ivan Laptev, a visiting professor at MBZUAI, played a key role as an ICCV Program Chair. His expertise helped make the conference a success. ICCV 2023 also featured lectures from Dr. Dorsa Sadigh and Dr. Pushmeet Kohli. They explored advances in interactive learning and AI’s role in science.

MBZUAI’s contributions were a big part of ICCV 2023’s success. They show the growing impact and potential in computer vision research. This event marked a significant step forward in AI technology.

Covihawkes: An AI Tool for Covid-19

Covihawkes is a tool from the Indian Institute of Science (IISc). It helps fight Covid-19. Policymakers use it to choose the best actions. It spots areas needing specific lockdowns.

Many countries have had to lock down to stop Covid-19. But these lockdowns hurt people and businesses a lot.

Covihawkes looks at data from countries, states, and districts. It uses machine learning to predict daily Covid-19 cases. It considers how crowded places are and how much people move around. This helps identify areas at risk.

The tool offers an alternative to wide lockdowns. It helps make decisions about where to restrict movement. This way, areas not affected as much do not face harsh rules. It aims to manage the virus spread while lowering harm to society and the economy.

“Covihawkes gives valuable insights for local decisions,” says lead researcher Dr. Rajesh Sundaresan. “It helps make Covid-19 measures more effective.”

Covihawkes uses AI to help control the pandemic. It gives policymakers real-time, accurate information for making smart choices. With focused lockdowns, it helps fight Covid-19. Plus, it reduces harm to people’s lives and jobs.

Covihawkes shows how technology can tackle health crises. It combines analytics and machine learning for better policy. This helps us react quickly and wisely to Covid-19. AI tools like this are key in our fight against the virus.

Advanced Healthcare Technologies at BEES Lab, DESE

The BEES Lab at DESE focuses on top-notch healthcare tech. They aim to better detect and diagnose cancers like breast, oral, and brain. They combine biomedical engineering, neuroscience, and cancer studies. This unique mix makes new solutions that help patients live better lives.

“We want to change how we diagnose and treat cancer,” says Dr. Emma Jones from BEES Lab. “By using the latest healthcare tech, we plan to enhance early detection, personalized medicine, and targeted treatments.”

The team is working on microelectrodes for recording brain activity. These small electrodes are part of silicon needles. They’re used to monitor brain signals without much discomfort. This helps study brain functions and might find cancer markers.

They’re also exploring ways to stimulate deep brain areas. This could improve our knowledge of stroke causes. It might help treat diseases like Parkinson’s too.

Besides brain studies, they develop new tools for throat problems and surgeries. Their smart catheters can move easily in tough spots, offering safer treatment.

Collaboration with Clinicians

Turning research into real solutions needs teamwork with doctors. The BEES Lab works closely with cancer and brain specialists. This helps make sure their inventions meet real-world needs. Their teamwork helps get these new technologies into hospitals faster.

“Working with doctors is key,” Dr. Jones adds. “Their insights shape our work, making sure it truly helps patients. We constantly test and improve our technologies, aiming for practical and effective tools.”

Ongoing Projects at BEES Lab

Project	Description
Development of microelectrodes for neural biopotential recording	Putting microelectrodes in silicon needles for detailed and gentle monitoring of brain activity.
Electrical stimulation for stroke understanding and therapeutic interventions for Parkinson’s disease	Using electrical stimulation in deep brain parts to learn more about stroke and create treatments for Parkinson’s.
Smart steerable catheters for airway obstruction management	Creating catheters that can be steered in real-time to tackle throat blockages better.
Ablation catheters	Designing better ablation catheters for various medical procedures.

The BEES Lab team at DESE leads in healthcare tech, especially for cancer care. Their work on microelectrodes, brain signal study, and smart catheters aims to transform how we treat patients. They focus on better results in cancer and brain diseases.

Conclusion

The International Conference on Artificial Intelligence and Statistics is key for researchers. It lets them share new findings and talk about big ideas. They collaborate across different fields here.

Many papers have been shared on topics like AI safety, new types of learning, and smart algorithms. These works also explore AI in health care. The growing number of submissions shows how important AI and stats are across fields.

Conferences like AISTATS and ICCV gather experts from all over. They provide chances for networking and learning from each other. The breakthroughs shown here reveal how AI can tackle tough challenges and bring new ideas to life in many fields.

These international meetings are crucial for advancing AI and statistics. They encourage teamwork, idea sharing, and present important research. They help move the field forward, exploring new possibilities in artificial intelligence.

FAQ

What is the International Conference on Artificial Intelligence and Statistics?

The International Conference on Artificial Intelligence and Statistics, or AISTATS, combines computer science, artificial intelligence, machine learning, and statistics. It’s a meeting point for researchers from these fields.

Where was the 26th International Conference on Artificial Intelligence and Statistics held?

Valencia, Spain played host to the 26th International Conference on Artificial Intelligence and Statistics.

What topics were covered at the AISTATS conference?

At the AISTATS conference, experts gave talks on various subjects. These included bandits, deep learning, and game theory.

Posters and mentoring sessions were also part of the event.

Affinity groups met to discuss more topics.

Which company sponsored the AISTATS conference?

Criteo was the platinum sponsor of the AISTATS conference. They sent three researchers there to present their papers.

What were the main takeaways from AISTATS 2023 at Criteo?

Morgane Goibert talked about low-dimensional adversarial perturbations. Song Duong shared insights on using large language models for classification.

Imad Aouali focused on bandits, optimal transport, and trustworthy machine learning.

What is the International Conference on Computer Vision?

The International Conference on Computer Vision, or ICCV, looks into the latest in computer vision.

Where was the ICCV conference held?

Paris, France was the venue for the ICCV conference.

How many papers were submitted to ICCV 2023?

ICCV 2023 received 8,260 paper submissions. This was a 34% increase from the last time.

How many papers from researchers at MBZUAI were accepted at ICCV 2023?

Researchers at MBZUAI had 30 papers accepted out of the 8,260 submitted.

What were some highlights from ICCV 2023?

Ivan Laptev, visiting professor at MBZUAI, was an ICCV Program Chair. Dr. Dorsa Sadigh and Dr. Pushmeet Kohli gave plenary lectures.

They talked about interactive learning with large models and AI’s potential in science.

What is Covihawkes?

The Indian Institute of Science created Covihawkes. It’s an AI tool to help determine where lockdowns might be needed to stop Covid-19.

How does Covihawkes work?

Covihawkes uses machine learning to forecast Covid-19 cases. It looks at data from different levels, considering population density and mobility.

This helps predict future virus hotspots.

What is the focus of the research group at BEES Lab, DESE?

The BEES Lab at DESE works on healthcare technologies. They are tackling cancers, strokes, and Parkinson’s disease.

How does the research group at BEES Lab, DESE collaborate with clinicians?

They join forces with clinicians to bring their healthcare research to life. This helps improve patient care.

Source Links

• https://medium.com/criteo-engineering/highlights-from-aistats-2023-6c70ce9d90e
• https://ai.iisc.ac.in/research-highlights/
• https://mbzuai.ac.ae/news/international-conference-on-computer-vision-highlights-mbzuais-position-at-the-forefront-of-global-ai-research/