BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Olle Häggström (Chalmers University of Technology\, Sweden)
DTSTART:20230821T140000Z
DTEND:20230821T150000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/1
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/1/">AI risk and AI alignment</a>\nby Olle Häggström (Chalmers Univers
 ity of Technology\, Sweden) as part of Mapping the Interdisciplinary Horiz
 ons of AI: Safety\, Functional Programming\, Information Geometry\n\n\nAbs
 tract\nThe planetary dominance over other species that humanity has attain
 ed has very little to do with muscular strength and physical endurance: it
  is all about intelligence. This makes the present moment in history\, whe
 n we are automating intelligence and handing over this crucial skill to ma
 chines\, the most important ever. The research area that has become known 
 as AI alignment deals with how to make sure that the first superintelligen
 t machines have goals and values that are sufficiently aligned with ours a
 nd that sufficiently prioritize human flourishing. This needs to succeed\,
  because otherwise we face existential catastrophe. In these lectures I wi
 ll outline key challenges in AI alignment\, what is being done to solve th
 em\, and how all this relates to the breakneck speed at which AI is presen
 tly advancing.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/1/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Olle Häggström (Chalmers University of Technology\, Sweden)
DTSTART:20230821T153000Z
DTEND:20230821T163000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/2
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/2/">AI risk and AI alignment</a>\nby Olle Häggström (Chalmers Univers
 ity of Technology\, Sweden) as part of Mapping the Interdisciplinary Horiz
 ons of AI: Safety\, Functional Programming\, Information Geometry\n\n\nAbs
 tract\nThe planetary dominance over other species that humanity has attain
 ed has very little to do with muscular strength and physical endurance: it
  is all about intelligence. This makes the present moment in history\, whe
 n we are automating intelligence and handing over this crucial skill to ma
 chines\, the most important ever. The research area that has become known 
 as AI alignment deals with how to make sure that the first superintelligen
 t machines have goals and values that are sufficiently aligned with ours a
 nd that sufficiently prioritize human flourishing. This needs to succeed\,
  because otherwise we face existential catastrophe. In these lectures I wi
 ll outline key challenges in AI alignment\, what is being done to solve th
 em\, and how all this relates to the breakneck speed at which AI is presen
 tly advancing.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/2/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Olle Häggström (Chalmers University of Technology\, Sweden)
DTSTART:20230822T140000Z
DTEND:20230822T150000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/3
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/3/">AI risk and AI alignment</a>\nby Olle Häggström (Chalmers Univers
 ity of Technology\, Sweden) as part of Mapping the Interdisciplinary Horiz
 ons of AI: Safety\, Functional Programming\, Information Geometry\n\n\nAbs
 tract\nThe planetary dominance over other species that humanity has attain
 ed has very little to do with muscular strength and physical endurance: it
  is all about intelligence. This makes the present moment in history\, whe
 n we are automating intelligence and handing over this crucial skill to ma
 chines\, the most important ever. The research area that has become known 
 as AI alignment deals with how to make sure that the first superintelligen
 t machines have goals and values that are sufficiently aligned with ours a
 nd that sufficiently prioritize human flourishing. This needs to succeed\,
  because otherwise we face existential catastrophe. In these lectures I wi
 ll outline key challenges in AI alignment\, what is being done to solve th
 em\, and how all this relates to the breakneck speed at which AI is presen
 tly advancing.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/3/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Noémie  Combe (Max Planck Institute for Mathematics in Sciences\,
  Germany)
DTSTART:20230822T153000Z
DTEND:20230822T173000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/4
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/4/">Exploring Information Geometry: Recent Advances and Connections to 
 Topological Field Theory</a>\nby Noémie  Combe (Max Planck Institute for 
 Mathematics in Sciences\, Germany) as part of Mapping the Interdisciplinar
 y Horizons of AI: Safety\, Functional Programming\, Information Geometry\n
 \n\nAbstract\nWith the rapid progress of machine learning\, artificial int
 elligence and data sciences\, the topic of information geometry is an impo
 rtant domain of research. We aim at introducing the topic of information g
 eometry\, as well as presenting some recent progress in this domain. Diffe
 rential geometry and algebraic aspects shall be developed. The new tight r
 elation between the information geometry and topological field theory will
  be discussed.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/4/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Noémie  Combe (Max Planck Institute for Mathematics in Sciences\,
  Germany)
DTSTART:20230824T140000Z
DTEND:20230824T160000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/5
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/5/">Exploring Information Geometry: Recent Advances and Connections to 
 Topological Field Theory</a>\nby Noémie  Combe (Max Planck Institute for 
 Mathematics in Sciences\, Germany) as part of Mapping the Interdisciplinar
 y Horizons of AI: Safety\, Functional Programming\, Information Geometry\n
 \n\nAbstract\nWith the rapid progress of machine learning\, artificial int
 elligence and data sciences\, the topic of information geometry is an impo
 rtant domain of research. We aim at introducing the topic of information g
 eometry\, as well as presenting some recent progress in this domain. Diffe
 rential geometry and algebraic aspects shall be developed. The new tight r
 elation between the information geometry and topological field theory will
  be discussed.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/5/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Frédéric Barbaresco (Thales Land and Air Systems\, France)
DTSTART:20230828T140000Z
DTEND:20230828T150000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/6
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/6/">Symplectic Foliation Structures of Information Geometry for Lie Gro
 ups Machine Learning</a>\nby Frédéric Barbaresco (Thales Land and Air Sy
 stems\, France) as part of Mapping the Interdisciplinary Horizons of AI: S
 afety\, Functional Programming\, Information Geometry\n\n\nAbstract\nWe pr
 esent a new symplectic model of Information Geometry  based on Jean-Marie 
 Souriau's Lie Groups Thermodynamics. Souriau model was initially described
  in chapter IV “Statistical Mechanics” of his book “Structure of dyn
 amical systems” published in 1969. This model gives a  purely geometric 
 characterization of Entropy\, which appears as an invariant Casimir functi
 on in coadjoint representation\, characterized by Poisson cohomology. Sour
 iau has proved that we can associate a symplectic manifold to coadjoint or
 bits of a Lie group by the KKS 2-form (Kirillov\, Kostant\, Souriau 2-form
 ) in the affine case (affine model of coadjoint operator equivariance via 
 Souriau's cocycle)\, that we have identified with Koszul-Fisher metric fro
 m Information Geometry. Souriau established the generalized Gibbs density 
 covariant under the action of the Lie group. The dual space of the Lie alg
 ebra foliates into coadjoint orbits that are also the Entropy level sets t
 hat could be interpreted in the framework of Thermodynamics by the fact th
 at dynamics on these symplectic leaves are non-dissipative\, whereas trans
 versal dynamics\, given by Poisson transverse structure\, are dissipative.
  We will finally introduce Gaussian distribution on the space of Symmetric
  Positive Definite (SPD) matrices\, through Souriau's covariant Gibbs dens
 ity by considering this space as the pure imaginary axis of the homogeneou
 s Siegel upper half space where Sp(2n\,R)/U(n) acts transitively. We will 
 also consider Gibbs density for Siegel Disk where SU(n\,n)/S(U(n)xU(n)) ac
 ts transitively. Gauss density of SPD matrices is then computed through So
 uriau's moment map and coadjoint orbits. Souriau’s Lie Groups Thermodyna
 mics model will be further explored in European COST network CaLISTA and E
 uropean HORIZON-MSCA project CaLIGOLA.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/6/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Frédéric Barbaresco (Thales Land and Air Systems\, France)
DTSTART:20230828T153000Z
DTEND:20230828T163000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/7
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/7/">Symplectic Foliation Structures of Information Geometry for Lie Gro
 ups Machine Learning</a>\nby Frédéric Barbaresco (Thales Land and Air Sy
 stems\, France) as part of Mapping the Interdisciplinary Horizons of AI: S
 afety\, Functional Programming\, Information Geometry\n\n\nAbstract\nWe pr
 esent a new symplectic model of Information Geometry  based on Jean-Marie 
 Souriau's Lie Groups Thermodynamics. Souriau model was initially described
  in chapter IV “Statistical Mechanics” of his book “Structure of dyn
 amical systems” published in 1969. This model gives a  purely geometric 
 characterization of Entropy\, which appears as an invariant Casimir functi
 on in coadjoint representation\, characterized by Poisson cohomology. Sour
 iau has proved that we can associate a symplectic manifold to coadjoint or
 bits of a Lie group by the KKS 2-form (Kirillov\, Kostant\, Souriau 2-form
 ) in the affine case (affine model of coadjoint operator equivariance via 
 Souriau's cocycle)\, that we have identified with Koszul-Fisher metric fro
 m Information Geometry. Souriau established the generalized Gibbs density 
 covariant under the action of the Lie group. The dual space of the Lie alg
 ebra foliates into coadjoint orbits that are also the Entropy level sets t
 hat could be interpreted in the framework of Thermodynamics by the fact th
 at dynamics on these symplectic leaves are non-dissipative\, whereas trans
 versal dynamics\, given by Poisson transverse structure\, are dissipative.
  We will finally introduce Gaussian distribution on the space of Symmetric
  Positive Definite (SPD) matrices\, through Souriau's covariant Gibbs dens
 ity by considering this space as the pure imaginary axis of the homogeneou
 s Siegel upper half space where Sp(2n\,R)/U(n) acts transitively. We will 
 also consider Gibbs density for Siegel Disk where SU(n\,n)/S(U(n)xU(n)) ac
 ts transitively. Gauss density of SPD matrices is then computed through So
 uriau's moment map and coadjoint orbits. Souriau’s Lie Groups Thermodyna
 mics model will be further explored in European COST network CaLISTA and E
 uropean HORIZON-MSCA project CaLIGOLA.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/7/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Seth Baum (Global Catastrophic Risk Institute\, University of Camb
 ridge\, USA)
DTSTART:20230829T140000Z
DTEND:20230829T150000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/10
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/10/">Social and Philosophical Dimensions of AI Alignment</a>\nby Seth B
 aum (Global Catastrophic Risk Institute\, University of Cambridge\, USA) a
 s part of Mapping the Interdisciplinary Horizons of AI: Safety\, Functiona
 l Programming\, Information Geometry\n\n\nAbstract\nSuccessful AI alignmen
 t requires three steps: (1) concept(s) for what\, if anything\, to align A
 I to\; (2) design(s) for how to implement the concept(s) in one or more AI
  systems\; and (3) the usage\, by human developers of AI systems\, of thes
 e design(s) instead of other\, less suitable design(s). Step (1) is mainly
  a matter of moral philosophy\; step (2) is mainly a matter of computer sc
 ience\; and step (3) is mainly a matter of social science and politics. My
  lecture(s) will cover steps (1) and (3).\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/10/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Seth Baum (Global Catastrophic Risk Institute\, University of Camb
 ridge\, USA)
DTSTART:20230829T153000Z
DTEND:20230829T163000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/11
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/11/">Social and Philosophical Dimensions of AI Alignment</a>\nby Seth B
 aum (Global Catastrophic Risk Institute\, University of Cambridge\, USA) a
 s part of Mapping the Interdisciplinary Horizons of AI: Safety\, Functiona
 l Programming\, Information Geometry\n\n\nAbstract\nSuccessful AI alignmen
 t requires three steps: (1) concept(s) for what\, if anything\, to align A
 I to\; (2) design(s) for how to implement the concept(s) in one or more AI
  systems\; and (3) the usage\, by human developers of AI systems\, of thes
 e design(s) instead of other\, less suitable design(s). Step (1) is mainly
  a matter of moral philosophy\; step (2) is mainly a matter of computer sc
 ience\; and step (3) is mainly a matter of social science and politics. My
  lecture(s) will cover steps (1) and (3).\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/11/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Patrik Jansson (Chalmers University of Technology\, Sweden)
DTSTART:20230830T140000Z
DTEND:20230830T150000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/14
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/14/">Domain-Specific Languages of Mathematics</a>\nby Patrik Jansson (C
 halmers University of Technology\, Sweden) as part of Mapping the Interdis
 ciplinary Horizons of AI: Safety\, Functional Programming\, Information Ge
 ometry\n\n\nAbstract\nThe main idea behind this minicoure is to encourage 
 the students to approach mathematical domains from a functional programmin
 g perspective. We will learn about the language Haskell\; identify the mai
 n functions and types involved\; introduce calculational proofs\; pay atte
 ntion to the syntax of mathematical  expressions\; and\, finally\, to orga
 nize the resulting functions and types in domain-specific languages.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/14/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Patrik Jansson (Chalmers University of Technology\, Sweden)
DTSTART:20230830T153000Z
DTEND:20230830T163000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/15
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/15/">Domain-Specific Languages of Mathematics</a>\nby Patrik Jansson (C
 halmers University of Technology\, Sweden) as part of Mapping the Interdis
 ciplinary Horizons of AI: Safety\, Functional Programming\, Information Ge
 ometry\n\n\nAbstract\nThe main idea behind this minicoure is to encourage 
 the students to approach mathematical domains from a functional programmin
 g perspective. We will learn about the language Haskell\; identify the mai
 n functions and types involved\; introduce calculational proofs\; pay atte
 ntion to the syntax of mathematical  expressions\; and\, finally\, to orga
 nize the resulting functions and types in domain-specific languages.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/15/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Patrik Jansson (Chalmers University of Technology\, Sweden)
DTSTART:20230831T140000Z
DTEND:20230831T150000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/16
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/16/">Domain-Specific Languages of Mathematics</a>\nby Patrik Jansson (C
 halmers University of Technology\, Sweden) as part of Mapping the Interdis
 ciplinary Horizons of AI: Safety\, Functional Programming\, Information Ge
 ometry\n\n\nAbstract\nThe main idea behind this minicoure is to encourage 
 the students to approach mathematical domains from a functional programmin
 g perspective. We will learn about the language Haskell\; identify the mai
 n functions and types involved\; introduce calculational proofs\; pay atte
 ntion to the syntax of mathematical  expressions\; and\, finally\, to orga
 nize the resulting functions and types in domain-specific languages.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/16/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Kazuki Okamura (Shizuoka University\, Japan)
DTSTART:20230831T153000Z
DTEND:20230831T163000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/17
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/17/">Metrization of powers of the Jensen-Shannon divergence [part of Wo
 rkshop]</a>\nby Kazuki Okamura (Shizuoka University\, Japan) as part of Ma
 pping the Interdisciplinary Horizons of AI: Safety\, Functional Programmin
 g\, Information Geometry\n\n\nAbstract\nMetrization of statistical diverge
 nces is useful in both the theoretical and practical aspects. \nConsiderin
 g the fractional powers of statistical divergences is one way to obtain me
 trics associated with divergences. \nWith this motivation\, Osán\, Bussan
 dri\, and Lamberti (2018) considered metrizations for the fractional power
 s of the Jensen-Shannon divergences between multinomial distributions and 
 gave an open problem. \nWe give an affirmative answer to their conjecture.
  \nOur method is also applicable to fractional powers of f-divergences bet
 ween the Cauchy distribution.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/17/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Dmitri Alekseevsky (University of Hradec Králové\, Czech Republi
 c)
DTSTART:20230823T140000Z
DTEND:20230823T150000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/18
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/18/">Neurogeometry of Vision and Information Geometry of Homogeneous Co
 nvex Cones</a>\nby Dmitri Alekseevsky (University of Hradec Králové\, Cz
 ech Republic) as part of Mapping the Interdisciplinary Horizons of AI: Saf
 ety\, Functional Programming\, Information Geometry\n\n\nAbstract\nThese l
 ectures will provide a comprehensive overview of information processing. W
 e will start with information processing in early vision in static and geo
 metric models of the primary visual cortex. Then\, we will explore informa
 tion processing in vision in dynamics. Finally\, we gonna talk about the i
 nformation geometry of Chentsov-Amari and homogeneous convex cones.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/18/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Dmitri Alekseevsky (University of Hradec Králové\, Czech Republi
 c)
DTSTART:20230823T153000Z
DTEND:20230823T163000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/19
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/19/">Neurogeometry of Vision and Information Geometry of Homogeneous Co
 nvex Cones</a>\nby Dmitri Alekseevsky (University of Hradec Králové\, Cz
 ech Republic) as part of Mapping the Interdisciplinary Horizons of AI: Saf
 ety\, Functional Programming\, Information Geometry\n\n\nAbstract\nThese l
 ectures will provide a comprehensive overview of information processing. W
 e will start with information processing in early vision in static and geo
 metric models of the primary visual cortex. Then\, we will explore informa
 tion processing in vision in dynamics. Finally\, we gonna talk about the i
 nformation geometry of Chentsov-Amari and homogeneous convex cones.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/19/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Frank Nielsen (Sony Computer Science Laboratories\, Japan)
DTSTART:20230829T070000Z
DTEND:20230829T090000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/20
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/20/">Introduction to Information Geometry\, Recent Advances\, and Appli
 cations</a>\nby Frank Nielsen (Sony Computer Science Laboratories\, Japan)
  as part of Mapping the Interdisciplinary Horizons of AI: Safety\, Functio
 nal Programming\, Information Geometry\n\nAbstract: TBA\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/20/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Michael Roop (Chalmers University of Technology\, Sweden)
DTSTART:20230831T150000Z
DTEND:20230831T153000Z
DTSTAMP:20260404T095454Z
UID:Wisla23/21
DESCRIPTION:Title: <a href="https://stable.researchseminars.org/talk/Wisla
 23/21/">Lie-Poisson methods for incompressible magnetohydrodynamics on the
  sphere [part of Workshop]</a>\nby Michael Roop (Chalmers University of Te
 chnology\, Sweden) as part of Mapping the Interdisciplinary Horizons of AI
 : Safety\, Functional Programming\, Information Geometry\n\n\nAbstract\nWe
  present a novel structure preserving numerical method for Lie-Poisson sys
 tems on the dual of semidirect product Lie algebras. The method preserves 
 the underlying geometry\, namely the Lie-Poisson structure and all the Cas
 imirs\, and nearly preserves the Hamiltonian function. We illustrate the m
 ethod on two models describing the motion of magnetic fluids\, the incompr
 essible magnetohydrodynamics equations and the Alfvén wave turbulence equ
 ations. For the latter case\, we reveal the formation of large scale quasi
 -periodic vortex blob dynamics.\n
LOCATION:https://stable.researchseminars.org/talk/Wisla23/21/
END:VEVENT
END:VCALENDAR
