Over the last few decades, developments in computing power, algorithms and, more recently, artificial intelligence, have continually enhanced the realism of numerical simulations of molecular systems. This workshops is motivated by recent advances in the Density functional theory based simulation of molecular systems in electronic excited states [e.g. 1-4, non exaustively] and/or those incorporating exotic particles such as muons or positrons, or quantum protons [5-8]. 

Meanwhile, numerical simulations also require interpretive models, which are essential for reducing the complexity of simulation data, synthesizing data according to shared theoretical frameworks, and eventually generalizing or predicting trends to other molecular systems. Theoretical concepts also facilitate experiment-theory interaction by providing shared comprehensive languages. Conceptual DFT (CDFT), introduced by Parr in the '80s for electronic systems in the ground state, has enabled key concepts from experimental chemistry, such as Hard and Soft Acids and Basis theory (HSBA), to be recovered in DFT [9]. DFT descriptors are defined as first or second derivatives of energy or electron density. These include chemical potential or electron affinity, chemical potential or hardness, Fukui reactivity functions or dual descriptors. Another approach is information theoretic approach (IPA) that uses concepts like Shannon or Fisher entropies, to quantify for example electro- and nucleophilicity or regioselectivity [10].  We  also mention direct analyses (DA) of electron density or derivatives of. Non-covalent interaction [11],   ultrastrong interaction [12] indicators have been derived. Topological analyses of dedicated scalar functions  [22] are other examples of DA. Only recently have researchers sought to extend CDFT/IPA/DTA to excited electronic states. Tognetti, Morell et al. extended the dual descriptor by using densities of excited electronic states to mimic the reactive partners [13]. Ayers, Liu et al. evaluated  polarizabilities of  molecules in their  excited states [14]. 

Moving to the realm of non-adiabatic molecular simulations, several groups coupled TD-DFT (within Real-Time [15] or Linear Response formalisms) to Ehrenfest, Surface Hopping or exact factorization schemes [1-4]. Multi-component DFT, which account for nuclear quantum effects or composite systems, are also booming these last years [5-8,16] 

Few studies have attempted to access CDFT/IPA/DTA descriptors along non-adiabatic dynamics, while they essentially await to de devised for non-purely electronic systems. Chattaraj and colleagues proposed a Quantum Fluid TDDFT [17]. Paul et al [18] followed chemical reactivity along TD-DFT surface hopping trajectories.  Pilmé and de la Lande extended ELF and QTAIM topological analyses to time-dependent densities [19]. Engel and Schrüger recently explored the use of Shannon Entropies and Mutual Information in ab initio MD simulations [20,21]. However, we note that these attempts are few, especially when compared with the success of interpretative approaches for the ground state.

Overall, our conviction is that while the community would greatly benefit from CDFT/ITA/DTA concepts for dynamics in excited states and multicomponent systems, there are still methodological and technical hurdles to overcome before they can be widely adopted. The aim of our workshop proposal is to bring together scientific communities to think together to move beyond the state of the art.