Nancy Xu

Advances in machine learning, coupled with rapidly growing genome sequencing and molecular profiling datasets, are catalyzing progress in genomics1. In particular, predictive machine learning models, which are mathematical functions trained to map input data to output values, have found widespread usage. Prominent examples include calling variants from whole-genome sequencing data2,3, estimating CRISPR guide activity4,5 and predicting molecular phenotypes, including transcription factor… Show more

Advances in machine learning, coupled with rapidly growing genome sequencing and molecular profiling datasets, are catalyzing progress in genomics1. In particular, predictive machine learning models, which are mathematical functions trained to map input data to output values, have found widespread usage. Prominent examples include calling variants from whole-genome sequencing data2,3, estimating CRISPR guide activity4,5 and predicting molecular phenotypes, including transcription factor binding, chromatin accessibility and splicing efficiency, from DNA sequence1,6,7,8,9,10,11. Once trained, these models can be probed in silico to infer quantitative relationships between diverse genomic data modalities, enabling several key applications such as the interpretation of functional genetic variants and rational design of synthetic genes. Show less

Groups and their derivative properties arise naturally in models of the world — bridging
intuition for graph symmetries and other mathematical structures. This presentation will
explore the question of how we can derive the order of a group given a set of its generators.
We show an algorithmic result by Furst-Hopcroft-Luks for finding the order of a symmetric
group subgroup in polynomial time.

This research explores deep Q-learning for autonomous driving in The Open Racing Car Simulator (TORCS). Using the TensorFlow and Keras software frameworks, we train fully-connected deep neural networks that are able to autonomously drive across a diverse range of track geometries. An initial proof-of-concept of classical Q-learning was implemented in Flappy Bird. A reward function promoting longitudinal velocity while penalizing transverse velocity and divergence from the track center is used… Show more

This research explores deep Q-learning for autonomous driving in The Open Racing Car Simulator (TORCS). Using the TensorFlow and Keras software frameworks, we train fully-connected deep neural networks that are able to autonomously drive across a diverse range of track geometries. An initial proof-of-concept of classical Q-learning was implemented in Flappy Bird. A reward function promoting longitudinal velocity while penalizing transverse velocity and divergence from the track center is used to train the agent. To validate learning, the research analyzes the reward function parameters of the models over two validation tracks and qualitatively assesses driving stability. A video of the learned agent driving in TORCS can be found online: https://www. dropbox. com/sh/b4623soznsjmp12/AAA1UD8_oaa94FgFC6eyxReya? dl= 0. Show less

Grounding natural language instructions on the web to perform previously unseen tasks enables accessibility and automation. We introduce a task and dataset to train AI agents from open-domain, step-by-step instructions originally written for people. We build RUSS (Rapid Universal Support Service) to tackle this problem. RUSS consists of two models: First, a BERT-LSTM with pointers parses instructions to ThingTalk, a domain-specific language we design for grounding natural language on the web��� Show more

Grounding natural language instructions on the web to perform previously unseen tasks enables accessibility and automation. We introduce a task and dataset to train AI agents from open-domain, step-by-step instructions originally written for people. We build RUSS (Rapid Universal Support Service) to tackle this problem. RUSS consists of two models: First, a BERT-LSTM with pointers parses instructions to ThingTalk, a domain-specific language we design for grounding natural language on the web. Then, a grounding model retrieves the unique IDs of any webpage elements requested in ThingTalk. RUSS may interact with the user through a dialogue (e.g. ask for an address) or execute a web operation (e.g. click a button) inside the web runtime. To augment training, we synthesize natural language instructions mapped to ThingTalk. Our dataset consists of 80 different customer service problems from help websites, with a total of 741 step-by-step instructions and their corresponding actions. RUSS achieves 76.7% end-to-end accuracy predicting agent actions from single instructions. It outperforms state-of-the-art models that directly map instructions to actions without ThingTalk. Our user study shows that RUSS is preferred by actual users over web navigation. Show less

This paper analyzes the mathematical foundations of Simpson's Paradox and proposes two distinct probabilistic and causal perspectives for understanding the phenomenon in practical application. Our research investigates data from observational studies and clinical trials on antibiotic prophylaxis for prevention of nosocomial infections -- one of the most common killers in hospitals worldwide.