*based on a call for papers solicitation at UCSB department of Geography …
*The three disruptive changes transforming science?
One-D AI revolution affecting media/society (fake news/AI-generated duplicity) but also science e.g., genetics/nanotech/transportation etc. This is a very generic broad expansion of AI as a support infrastructure to speed up information. However, not highly effective in novel solutions (NOT ready to replace innovators but good stochastic replicators e.g., writing 21st century Shakespearean novels!).
Three-D electromechanical-AI with hyper-powered ChatGPT4-type language model inspired mechanical actions – a revolution in manufacturing and transportation-automation.
Quantum Computing (RSA-type cryptography under threat – to affect commerce/internet security/information ecosystem); post-quantum cryptography using lattice-based constructions are promising! Lattice-based (spatial!) constructions may have potential applications in spatial cyber-security. I can think of applications in last-mile authentication protocols for drones that overcome quantum hacks on RSA type cryptography.
However, to use the air for transport will be challenging and may fail without a comprehensive psycho-social and psycho-political deprogramming of past transport – from horses/cars to drones (to replace roads!).
*Biggest impact of future disruptive science on GIScience
GISc should lead in:
- 3D micro-meso-macro simulation; network science theories built on top of Operations Research, as it is our fiduciary duty to minimize spatial waste.
- Automatic production of Bayesian (unbiased) posterior estimates over raw-data/likelihoods. No one should use raw data from political units to perform social/behavioral analysis – another step towards hard social science. Bayesian hierarchical networks to identify hidden chains and networks that underlies the socio-economic and built-form biases –a geo-ecological network (it includes the psycho-social and psycho-political).
- Realistic simulation of cities with fully loaded dynamic information layers. GISc provides an ecological platform for complexity analysis. No one can solve complex problems in their scientific silos.
- Finally, a long shot, but Quantum Computing may help solve NP-Hard problems with provable algorithms instead of heuristics e.g., redistricting. GISc can adopt QC now: e.g., use quantum entanglement to create predictive cartography (display only the info you need at scale).
- Replace desktop with Cloud-based/HP computing.
*What should we prioritize over the next decade?
- Spatial Stochastic Operations Research (Spatial Stochastic OR), like Computer Simulation: Transportation space-time simulation of autonomous ground vehicles (most cars will become autonomous; and drones with maximum level 5 automation). Stochastic Optimizations like Poylak averaging of 100 shortest paths trip for each vehicle in each lane every 1/10 of a second (maximize safety at present speeds). Dynamic worlds, like disaggregated pollution profiles by morning traffic near playgrounds can be simulated on the fly.
- Stochastic OR can be applied to other areas of GIS like climate models using network complexity
- Software tools for citizens science (GIS as a lithospheric inventory management system; redistricting tools, location-allocation tools; simulation tools especially for climate change and risk management).
- Spatial Stochastic OR can become part of citizen GISc (PPGIS) when Apple™ etc. develop haptic virtual reality SPATIAL products.
- A focus to utilize quantum computing for spatial optimization. One promising area is using lattice-based (NP-Hard) computing to authentic autonomous delivery systems.
- Mathematical statistics tools that utilize spatial-astronomy tools (Markovian, monte-carlo, correlation, entropy), spatial-econometrics tools, geography/social-science tools,
- Neural-net AI: GIScientists should prioritize use over development. AI is fast changing and require silo-like dedication to advance. The best case for spatial data scientists and GIScientists is to adopt the latest. ChatGPT 3.5 and 4 are worlds apart! So, no use spending time in development.
- Quantum Computing: Developing Quantum computing algorithms to solve spatial NP-Hard problems will be game-changing. It is surprisingly simple how some QC algorithms are being adopted since quantum entanglement allows for information storage (not retrieval) at exponentially smaller Qbits compared to bits. However, although Qbits are amazingly efficient at storage, retrieval collapses to only one state (deleting all other information). This is due to quantum entanglement, but can be avoided sometimes by using tricks in mathematics like using “quantum Fourier transforms” to break RSA type prime numbers factorization.
*GIS/Data Science curriculum in an age of disruption and scientific transformation?
Operations Research (OR) knowledge is key to mitigate inefficiencies especially due to climate change (do less with more). Spatial OR can play a massive role — from auto-focused use of room lighting – a spatial issue, to amazon deliveries that lowers our carbon footprint.
Spatial Simulation (includes stochastic OR): In healthcare, GISc inspired hospital traffic can monitor hospital-based infection propagation e.g., Nosocomial infections like MRSA kill 200K/year (US).
Meso-scale operations: My advanced GISc students create, for 2040, city-level (meso-micro) traffic simulation using level-5 autonomous drones – a probable future scenario!
Therefore, GISc involves the micro (3D architecture), the meso (regional e.g., epidemic trends) and macro (global: e.g., climate-change, macro-economic, ecologic etc.) spatial operations. Stochastic optimization is a complex area involving probabilistic optimization (e.g., using dynamic programming to solve traffic). In fact, stochastic optimization is now part of connected cars used by GM, Ford etc. GIScientists can become experts in autonomous trip movements.
GISc curriculum may include:
- GISc as a micro-meso-macro scale operating system for mimicking (digital twinning) the earth surface. This is not only a functional convenience or ontological representation of our world (and run models), but also to understand the deeper epistemic underpinnings of our society. From architects, to macro-economists, sociologists and poets, everyone can view the world through space and time with mouse clicks or gestures.
- GISc AS INVENTORY MANAGEMENT for Climate/morphological change/action.
- GISc AS A FORECASTER.
- GISc, as not just an efficient/positivist/capitalist endeavor but also, an epistemic, as well as an ontological/hermeneutic/phenomenological/humanistic one that cherishes society/environment/ecology through public and private action.
- Spatial philosophy (distance creates inefficiencies in market but can be optimized. From market economy to sustainable ecology solutions.
- Planar and non-planar topology from scratch (students build spatial hierarchies using graph-theoretic tools).
- Database (SQL-type single machines relational databases to distributed hierarchical ones)
- Simulation from scratch so that students can compute: e.g., a 8-9 AM pollution profile for a downtown school using Comprehensive Modal Emission Model (CMEM) emissions (my students do this in my 300 level undergrad environmental GISc classes).
- Location Science using integer programming/network flows: network partitioning, arc routing, network flow models, location-allocation, regional partition (anti-gerrymandering).
- Hierarchical regression using simple empirical as well as conditional autoregressive models based on random walks (Markovian) and computed using integrated nested Laplace Transformations (e.g., using R-INLA software).
Leave a Reply