WiDS Stockholm 2023 Conference

Drop in at any time after 8:30 to register for the event, pick up your name badge and start mingling with other attendees! Coffee and tea will be available.

What is Women in Data Science, who are the organizers, why are we doing this in our spare time and what else do we do than run this event? Join to find out!

The classification problems we work with at Tink are inherently subjective, where identical data points can mean different things to different users. This presents challenges for modelling and evaluation when compared to standard multiclass classification problems. This talk will be a deep dive into how we have approached evaluation and optimization of our models to balance between trying to find the true "correct" category for a specific user and trying to predict a category that is "not wrong" for the general population, and how to communicate the results to business stakeholders.

Data is often broken. Data analysts and scientists complain that data is broken. Data engineers complain that analysts and scientists complain. This talk aims to bridge the gap between domains by giving the data engineering perspective on some common reasons and solutions to broken data.

The topic this year is “Evaluation methods in Data Science and Machine Learning” - methods of our day-to-day work which are fundamental to understand and estimate the impact Data Science and Machine Learning have on the product, people and even the society. These short 6 minute talks will each highlight a specific method, a problem that this method helped you to solve and what was an evaluation scenario you used this method in. Topics might include but are not limited to:

• Impact evaluation (Causal inference and counterfactuals)
• Experimentation and A/B testing
• Offline machine learning model evaluation. In particular we are interested in simulation methods when a real-time or online evaluation is expensive or impossible to build.
• Evaluation methods for data and data analysis success
• Product strategy evaluation (North star KPI, data-driven product development)

Read more about this and reach out to us if you'd like to become a lightning talk speaker here: SIGN UP TO GIVE A LIGHTNING TALK

Lunch will be served at Epicenter for everyone who is attending in person. We will make sure there's a selection of vegetarian and vegan options, but if you have very specific dietary requirements, shoot us an email and we will see what we can do!

This talk will focus on the different flavours of analytics work in modern tech companies. While traditional product analytics centers around helping product managers better understand and improve product performance, analysts can also be directly involved in building customer-facing data products. While the underlying metrics and pipelines they build might differ, analysts play an integral part in both cases and this talk will cover lessons learned from working on both sides.

"The way you speak should empower you, not to hold you back in your work, your career, or your life." This talk is about how Astrid develops and scales Deep Learning and Generative technology to make spoken communication training accessible, effective, and affordable for everyone. Large AI models serve in our core system and live process users’ speech. They are accompanied by multiple ML and data services to formulate feedback, guidance, recommendations, and insights for users. All these need to serve users in a seamless and engaging experience.

Despite their fantastic achievements in fields such as computer vision and natural language processing, state-of-the-art deep learning approaches differ from human cognition in fundamental ways. While humans can learn new concepts from just a single or few examples, and effortlessly extrapolate new knowledge from concepts learned in other contexts, deep learning methods generally rely on large amounts of data for their learning. Moreover, while humans can make use of contextual knowledge of e.g. laws of nature and insights into how others reason, such information is generally hard to exploit in deep learning methods.Current deep learning approaches are indeed purposeful for a wide range of applications where there are large volumes of training data and/or well defined problem settings. However, models that learn in a more human-like manner have the potential to be more adaptable to new situations, be more data efficient and also more interpretable to humans - a desirable property e.g. for intelligence augmentation applications with a human in the loop, e.g. medical decision support systems or social robots.In this talk I will describe a number of projects in my group where we explore disentanglement, temporality, multimodality, and cause-effect representations to accomplish compositional, structured, and interpretable models of the world.

A recent blog post recently sparked some interesting debate in the data world. Are data teams really providing (enough) value at companies today, or is the fact that many data teams are "navigating the same problems of being misunderstood, left out, and asked to do the wrong work" an indication that the problem is the data teams and not everyone else? And how does the current marco-economic situation affect us? Will data teams be considered a cost center that needs to be trimmed down? Or is there still a bright future for data teams and if so, what does it look like? What can I, as an individual, do to make sure I'm creating clear and visible value through my work?

This panel will dive into these and other tough questions, with the aim of introspecting our field and suggesting ways to ensure your work stays relevant and valuable. We will also open up for questions from the audience!

Panelists:

• Ylva Lundegård, Data Engineer at EQT Partners
• Lena Sundin, Freelance Backend and Data Engineer
• Andreea Taylor, Staff Machine Learning Engineer at Voi

Moderator: Rebecka Storm, Co-founder at Peach Data

Continue the evening by mingling, networking and discussing with fellow data and ML people! Meet new people, discuss your takeaways from the talks of the day, learn about what others do in a relaxed and friendly environment. The mingle will be across the intersection at the headquarters of one of our sponsors, King.