 Nie jesteś zalogowany | zaloguj się |
| Nie jesteś zalogowany | zaloguj się |
2023-01-19, godz. 12:15, 5050
Michal Nauman (Uniwersytet Warszawski)
In this talk, we will discuss policy gradients with many action samples. We will investigate decompositions of policy gradient variance, as well as measure the variance reduction effect stemming form increasing the number of state and action samples used in estimation. Finally, we will compare vario...
2022-12-08, godz. 12:15, https://uw-edu-pl.zoom.us/j/97528316706
Spyros Mouselinos (Uniwersytet Warszawski)
A Simple, Yet Effective Approach to Finding Biases in Code Generation
Recently, scores of high-performing code generation systems have surfaced. As has become a popular choice in many domains, code generation is often approached using large language models as a core, trained under the masked or causal language modeling schema. This work shows that current code gene...
2022-11-03, godz. 12:15, 5050
Piotr Tempczyk (Uniwersytet Warszawski)
One Simple Trick to Fix Your Bayesian Neural Network
One of the most popular estimation methods in Bayesian neural networks (BNN) is mean-field variational inference (MFVI). In this work, we show that neural networks with ReLU activation function induce posteriors, that are hard to fit with MFVI. We provide a theoretical justification for this phenome...
2022-10-27, godz. 13:15, https://uw-edu-pl.zoom.us/j/97528316706
Patrik Reizinger (University of Tübingen)
Embrace the Gap: VAEs Perform Independent Mechanism Analysis
Variational autoencoders (VAEs) are a popular framework for modeling complex data distributions; they can be efficiently trained via variational inference by maximizing the evidence lower bound (ELBO), at the expense of a gap to the exact (log-)marginal likelihood. While VAEs are commonly used for r...
2022-06-09, godz. 12:15, https://us02web.zoom.us/j/83632849075
Shadi Shafighi (Uniwersytet Warszawski)
Tumoroscope: a probabilistic graphical model for mapping tumor clones in cancerous tissue
Tumor cell populations are highly heterogeneous and form clones with different genotypes. Geographically distinct parts of the tumor have different genetic and phenotypic compositions. Elucidating tumor heterogeneity is hampered by the fact that there is no techn...
2022-06-02, godz. 12:15, 3140
Sebastian Jaszczur (Uniwerystet Warszawski)
Conditional Computation in Transformers
Note: seminar will be in person with a follow-up lunch. Transformer architecture is widely used in Natural Language Processing to get state of the art results. Unfortunately, such model quality is usually only possible by using extremely large models, which require significant resources dur...
2022-05-26, godz. 12:15, https://us02web.zoom.us/j/83632849075
Nabil Kahouadji (Northeastern Illinois University)
In our retrospective study, we evaluate associations between neighborhood-level indicators of structural violence and colorectal adenoma using University of Illinois Health electronic medical record (EMR) data obtained from patients receiving screening colonoscopy between the year 2015 and ...
2022-04-07, godz. 12:15, 3140
Piotr Kozakowski (Uniwerystet Warszawski)
Recent works have shown the effectiveness of entropy regularization in Monte Carlo Tree Search (MCTS). In this presentation I will first introduce the framework of Maximum Entropy Reinforcement Learning and show how it can be applied to MCTS. Then I will present various variants of entropy regulariz...
2022-03-31, godz. 12:15, meet.google.com/ooi-zxye-dxa
Adam Izdebski (Uniwerystet Warszawski)
Generative Modelling with Optimization for Molecule Discovery
In recent years, discovering novel drug-like molecules become a common application of generative models. However, it is much harder to generate novel molecules that are at the same time optimized for being promising drug candidates. During the talk, I will give a snapshot of one of the many ...
2022-03-17, godz. 12:15, meet.google.com/ooi-zxye-dxa
Spyros Mouselinos (University of Warsaw)
Measuring CLEVRness: Black-box Testing of Visual Reasoning Models
How can we measure the reasoning capabilities of intelligence systems? Visual question answering provides a convenient framework for testing the model's abilities by interrogating the model through questions about the scene. However, despite scores of various visual QA datasets and architectures...
