Nexus of Biology and Computing

Slideshow Transcript

1. Slide 1: The nexus of biology and computing Small scale and complexity are forcing advances in computational methodologies Melanie Swan, Futurist MS Futures Group 415-505-4426 BCIG NIH melanie@melanieswan.com May 24, 2007 http://www.melanieswan.com http://futurememes.blogspot.com
2. Slide 2: Bio – Melanie Swan Educational background:  BA French & Economics, Georgetown University  MBA Finance & Accounting, Wharton, Univ. of Pennsylvania  Current course work in Physics & Computer Science  Professional experience  Futurist: speaker, researcher, business advisor  Hedge Fund Manager: Wall Street, proprietary  Current projects  OpenBasicResearch.org  del.icio.us for people  Issues in running Historical Simulations  Interests: science fiction, travel  BCIG May 24, 2007 2
3. Slide 3: Summary: Seven principles suggest future advances in computational methodologies Approaches to computation – approaches of parallelism 1. Architecture – modularity, simplicity and ubiquity of structure 2. Goals – broadly defined objectives to drive higher value results 3. Modulation mechanisms – information modulation 4. Prediction mechanisms – probabilistic models 5. Unconscious processing – unobtrusiveness computing 6. Multidisciplinarity 7. – adjacent discipline integration BCIG May 24, 2007 3
4. Slide 4: 1. Approaches to computation Traditional: Von Neumann  Linear  Current and future: non-Von Neumann  Cellular, tissue, systemic, holistic focus  Parallelism and multicores in hardware and software  DNA computing  Quantum computing  Genetic computing  Evo-devo: blend of bottom up emergence / top down design  Suggests biological and other approaches facilitating  parallelism are required for molecular scale computing BCIG May 24, 2007 4
5. Slide 5: 2. Architecture Conservation  Across simple and complex organisms  Across processes within one organism  Across time, evolution  Structure  Same loose administrative over-structures, diverse applications  Redundancy in architecture and process  Massively distributed individual agents  Suggests modularity, simplicity and ubiquity of underlying  structure BCIG May 24, 2007 5
6. Slide 6: 3. Goals Systemic, holistic Traditional, singular Clusters of functionality, One precise goal or   capability, redundancy outcome Loose process, many Tightly directed process   outcomes coupled to outcome Service paradigm Task paradigm   Focus on obtaining Exclusive focus on THE   useful information solution Suggests more broadly defined objectives drive higher  value results BCIG May 24, 2007 6
7. Slide 7: 4. Modulation mechanisms Short and long-term memory:  An implemented evaluation of the importance of information  Brain automatically modulates importance  Computing can better modulate information with attributes  signaling relevance, value, accuracy, etc. Repetition, time-based algorithms  Web 2.0 marks relevance and importance  Scientific Research 2.0 – digg for PubMed, RSS peer feeds,  collaborative research paper commenting and annotation Suggests much higher levels of information modulation  with relevance attributes BCIG May 24, 2007 7
8. Slide 8: 5. Prediction mechanisms Prediction is a strong biological mechanism  Explosion in predictive, probabilistic, statistical,  Bayesian papers and applications Numenta  Google  Key parameters of successful probabilistic model  implementation Large data corpus  Abstraction processes  Suggests greater development and application of  probabilistic models BCIG May 24, 2007 8
9. Slide 9: 6. Unconscious processing Brain processes mainly unconsciously  Some computer processing is “unconscious”  AI, virus scans, ajax websites  Other computer processing is very obvious  Memory, processing, storage  Heat, power, battery  Connectivity  Processing will become less conscious  Wearables, pen computing, visualization, simulation  Ubiquitous embedded chips, sensors, connectivity  Suggests a focus on less obtrusiveness computing  BCIG May 24, 2007 9
10. Slide 10: 7. Multidisciplinarity Cross-field collaboration and new area definition  Molecular cognition, molecular science of behavior  Systems biology  Quantitative measurement and mathematical analysis  Systems level studies: focus on quantitative aspects and  interactions among elements Need to standardize: an eigenvalue by any other name  Multidisciplinary cataloging of all biological information  E.O. Wilson Encyclopedia of Life  Suggests greater integration of adjacent disciplines in  pursuit of open research questions BCIG May 24, 2007 10
11. Slide 11: Summary: Seven principles suggest future advances in computational methodologies Approaches to computation – approaches of parallelism 1. Architecture – modularity, simplicity and ubiquity of structure 2. Goals – broadly defined objectives to drive higher value results 3. Modulation mechanisms – information modulation 4. Prediction mechanisms – probabilistic models 5. Unconscious processing – unobtrusiveness computing 6. Multidisciplinarity 7. – adjacent discipline integration BCIG May 24, 2007 11
12. Slide 12: Melanie Swan, Futurist Thank you MS Futures Group 415-505-4426 melanie@melanieswan.com http://www.melanieswan.com http://futurememes.blogspot.com