For the fifth year in a row, Harvard, MIT, Microsoft Research New England, and now the Broad Institute, are proud to collaborate with Stanford University to bring the Women in Data Science (WiDS) conference to Cambridge, Massachusetts.

This virtual, one-day technical conference will feature an all-female lineup of speakers from academia and industry, to talk about the latest data science-related research in a number of domains, to learn how leading-edge companies are leveraging data science for success, and to connect with potential mentors, collaborators, and others in the field.

WiDS Cambridge is an independent event that is organized by Harvard, MIT, Microsoft Research New England, and the Broad Institute, as part of the annual WiDS Worldwide conference organized by Stanford University and an estimated 150+ locations worldwide, which features outstanding women doing outstanding work in the field of data science. All genders are invited to attend all WiDS Worldwide conference events.

The post Women in Data Science Conference – 3/11 appeared first on Artificial Intelligence.

In celebration of International Women’s Day, women data scientists from around the world will gather together virtually on March 7 and 8, 2021, for presentations, workshops and networking at the Women in Data Science (WiDS) Worldwide Conference. The event offers women in data science, a chance to recognize their social, economic, cultural and political achievements.

“Data science is the new gold or the new oil,” WiDS co-founder Dr. Margot Gerritsen said. “Data-driven decision making is in all aspects of our lives. It’s important that all sorts of genders, cultures and backgrounds are represented in this development.”

Like many women working in fields still dominated by men, Dr. Margot Gerritsen remembers the tipping point that inspired her to co-found WiDS with Karen Matthys. Invited to attend a conference by a male colleague, she noted that the event featured no women keynote speakers. Her colleague offered a familiar explanation: There were so few women in the field that he couldn’t find one to present at the event.

Less than a year later, the first WiDS Worldwide Conference proved him wrong: Over 6,000 people from around the world joined the conference livestream. By 2020, over 30,000 participants in more than 50 countries tuned in.

According to conference organizer Judy Logan, the WiDS Worldwide Conference has been global since its start. The initial conference livestream’s popularity showed co-directors Gerritsen, Logan and Matthys the international demand for collaboration and networking among women data scientists around the world.

The virtual nature of the conference allowed women to connect with one another across borders without the elitist registration and travel costs of other conventions. The WiDS Worldwide Conference encourages women in data science to establish their own national and regional networks while participating in a broader global conversation. As advocates recognize, data science needs diverse perspectives to determine the future of data-driven decision making in a world where data collection and analysis are increasingly (and unevenly) penetrating every aspect of peoples’ lives.

The opening presentations for this year’s conference include a fireside chat with Nobel Laureate and physicist Andrea Ghez—only the fourth woman to win the Nobel Prize for Physics—and a presentation by Emily Fox, distinguished engineer (Apple) and professor (University of Washington) on health and machine learning. Additional presentations center the work and research of data scientists working around the world.

For example, Dr. Kalinda Griffiths, a scientist at the Centre for Big Data Research in Health at the University of New South Wales, will discuss her research addressing health disparities among indigenous peoples, including how they have been defined and how identifications are operationalized in government data collections.

Dr. Fatima Abu Salem, associate professor of computer science at the American University of Beirut (AUB), will discuss “Doing Data Science in Data Deserts,” a presentation that looks at how grassroots initiatives are facilitating data collection efforts in countries where data is less abundant and accessible.

This year’s conference features the fourth incarnation of WiDS datathon, an event in which participants are given a dataset and a problem for which they must use their ingenuity and data science skills to explore solutions. This year’s datathon reflects WIDS’ commitment to solving problems with global ramifications: Datathon participants have been asked to create models to determine whether patients have been diagnosed with a certain type of diabetes that might improve ICU treatment.

The WiDS Worldwide Conference is open to all interested in ensuring that the future of data science is diverse and inclusive. Information about registration fees and platform can be found on the WIDS website.

The post Women in Data Science Initiative Holds Global Conference to Celebrate International Women’s Day appeared first on Artificial Intelligence.

The 35th AAAI Conference on Artificial Intelligence (AAAI-21), held virtually this year, saw more than 9,000 paper submissions, of which, only 1,692 research papers made the cut.

The Association for the Advancement of Artificial Intelligence (AAAI) committee has announced the Best Paper and Runners Up awards. Let’s take a look at the papers that won the awards.

Best Papers

1| Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting

About: Informer is an efficient transformer-based model for Long Sequence Time-series Forecasting (LSTF). A team of researchers from UC Berkeley introduced this Transformer model to predict long sequences. Informer has three distinctive characteristics:

• A ProbSparse Self-attention mechanism, which achieves O(Llog L) in time complexity and memory usage, has comparable performance on sequences’ dependency alignment.
• The self-attention distilling highlights dominating attention by halving cascading layer input, and efficiently handles extreme long input sequences.
• The generative style decoder that predicts the long time-series sequences at one forward operation rather than step-by-step, which improves the inference speed of long-sequence predictions.

2| Exploration-Exploitation in Multi-Agent Learning: Catastrophe Theory Meets Game Theory

About: Exploration-exploitation is a powerful tool in multi-agent learning (MAL). A team of researchers from Singapore University of Technology studied a variant of stateless Q-learning, with softmax or Boltzmann exploration, also termed as Boltzmann Q-learning or smooth Q-learning (SQL). Boltzmann Q-learning is one of the most fundamental models of exploration-exploitation in MAS.

3| Mitigating Political Bias in Language Models through Reinforced Calibration 

About: Researchers from Dartmouth College, University of Texas and ProtagoLabs described metrics for measuring political bias in GPT-2 generation and proposed a reinforcement learning (RL) framework to reduce political biases in the generated text. Using rewards from word embeddings or a classifier, the RL framework guided the debiased generation without having access to the training data or requiring the model to be retrained. The researchers also proposed two bias metrics (indirect bias and direct bias) to quantify the political bias in language model generation.

Runners Up

1| Learning from eXtreme Bandit Feedback

About: Researchers from Amazon and UC Berkeley studied the problem of batch learning from bandit feedback in extremely large action spaces. They introduced a selective importance sampling estimator (sIS) operating in a significantly more favorable bias-variance regime. The sIS estimator is obtained by performing importance sampling on the conditional expectation of the reward concerning a small subset of actions for each instance.

2| Self-Attention Attribution: Interpreting Information Interactions Inside Transformer

About: Researchers from Microsoft and Beihang University proposed a self-attention attribution algorithm to interpret the information interactions inside the Transformer. As part of the research, the scientists first extracted the most salient dependencies in each layer to construct an attribution graph, which reveals the hierarchical interactions inside the Transformer. Next, they applied self attention attribution to identify the important attention head. Finally, they showed that the attribution results can be used as adversarial patterns to implement non-targeted attacks towards BERT.

3| Dual-Mandate Patrols: Multi-Armed Bandits for Green Security

About: Researchers from Harvard University and Carnegie Mellon University introduced LIZARD, an algorithm that accounts for decomposability of the reward function,  smoothness of the decomposed reward function across features, monotonicity of rewards as patrollers exert more effort, and availability of historical data. According to them, LIZARD leverages both decomposability and Lipschitz continuity simultaneously, bridging the gap between combinatorial and Lipschitz bandits.

The post Which Papers Won At 35th AAAI Conference On Artificial Intelligence? appeared first on Artificial Intelligence.