"Statistics at a Crossroads: Challenges and Opportunities in the Data Science Era" talks 2:00pm Monday #JSM2019 in CC-207 hub.ki/groups/statscr… hear from Bin Yu, David Banks, Dylan Small, Marianthi Markatou, David Madigan, Xuming He, and Michael Jordan @InstMathStat @CMU_Stats 
@InstMathStat @CMU_Stats Irrespective of whether you can attend or not, grand challenge ideas and reflections on the draft report are solicited at docs.google.com/forms/d/e/1FAI…
@InstMathStat @CMU_Stats Key takeaway #1: 1. The central role of practice. Today it is imperative for us to put practice at the center of our discipline with relevant computation and theory as supports. Research and education in statistics and data science must aim at solving real world problems.
@InstMathStat @CMU_Stats Key idea #2: Emphasis on impact. The profession needs to place more emphasis on the scientific and societal impact of statistical research in the evaluation of scholarship.
@InstMathStat @CMU_Stats Key idea #3. Research for better practice. For statistics research to effectively support science and real world data problem solving, it goes without saying that statistics research formulations have to reflect and capture the realities present in modern data problems.
@InstMathStat @CMU_Stats Key idea #4: Embracing grand challenges. The profession is ready to take on big research questions for the development of empirically proven statistical investigation processes, including problem formulation, data processing, & statistical and machine learning methods/algorithms
@InstMathStat @CMU_Stats #4 cont.: for the analysis of emerging data types, the development to relevant theory to support and advance such endeavors, & the development of computational platforms that account for various trade-offs in statistical efficiency, computation, communication, & storage costs.
Xuming He @UMichResearch kicks off the Statistics at a Crossroads late breaking session to a full house @ASA_SciPol #jsm2019
What would Tukey say? We need a cultural change. At least that’s what he said in 1962. #JSM2019
David Banks @DukeDataFest the future of Statistics lies in applications: astrostatistics, computational advertising, particle physics, driverless cars, satellite images and spandrels
David quoted @hadleywickham: a data scientist is a statistician who is useful. And then David slammed the qualifiers in stats from @DukeU and @NCState #JSM2019
Michael Jordan @UCBerkeley describes #Statistics as a problem solving culture and that we need to embrace being engineers (in a new subfield of engineering). How can we use data to make the world a better place? #JSM2019
Michael Jordan’s Grand Challenges: #hilbert for our time? #statistics often far too unambitious. These fourth generation problems are worth tackling. #JSM2019 #datascience
Marianthi Markatou reminds us of the 2004 statistical science report on the future of statistics and highlights opportunities in #PrecisionMedicine
Dylan Small #causal inference is at the heart of everyday life. New opportunities from “data exhaust” #JSM2019. Need to capitalize on multiple datasets eg high precision moderate bias @_MiguelHernan
Bin Yu @UCBerkeley argues for the importance of domain applications. How can advising models change to be more flexible? #JSM2019 need to write but higher quality papers.
Experimental design is dead (except for clinical trials and pragmatic trials). But causal inference is critical. #JSM2019
The world is full of unstructured data. How can we engage? Bin suggests we need to follow the problems. We too often define problems based on what we learned in textbooks and graduate school. #JSM2019
Michael Jordan: training in statistics might need to have less breadth but be able to learn new things @DebAtStat #JSM2019 #statsed
Roy Welsch what is the role of just in time education. Michael Jordan described undergraduate initiatives at @UCBerkeley to prepare for #workforce. Different challenges for leaders at the graduate level. Need new colleges for equivalent of #datascience. #JSM2019
How do we foster statistical practice? What structures and mechanisms can support exploration? How can we encourage? Money is really important (not just @NSF). We need to be more entrepreneurial.
David Banks: our publication and promotion systems are flawed. Education is difficult to change but needs to be rebuilt from ground level. #JSM2019 marianthi describes examples of mechanisms from #biostatistics @THISISBIOSTAT
No difference between statistical engineering and computer engineering. We need to think algorithmically and avoid thinking about arbitrary differences. Examples data provenance as a provably correct system. Michael Jordan #JSM2019
Great comment from @IyueSung: need more industry connections and involvement to really make change.
Key role that socialization plays in graduate school. How can we use this process to improve our programs?
• • •
Missing some Tweet in this thread? You can try to
force a refresh
