256869 research outputs found
Sort by
Exploring the Benefits of Iterative Retrieval-Augmented Generation for Risk Mitigation in LLM Responses
The correctness of provided information is essential for LLMs to be used in real life scenarios. However, they suffer from knowledge cut-offs as well as hallucinations. Retrieval-augmented generation (RAG) aims at solving these problems by providing on-demand, real-time information, usable as context for addressing queries. An issue with RAG based systems is the potential of low quality retrievals, that provide wrong or irrelevant context to the LLM. To mitigate this risk, we implemented Iter-RAG, an evolution of standard RAG which utilizes multi-iteration document retrieval based on missing information. Using Iter-RAG, the correctness of answers for the TriviaQA dataset is increased compared to standard RAG. Additionally, we implemented a chatbot that can correctly present information about an energy provider, admit when information is missing in its database, and refuse to answer adversarial queries, even when the requested information is present
Optimization of prospective circular economy in sewage sludge to biofuel production pathways via hydrothermal liquefaction using P-graph
Generalized Hydrodynamics: A Perspective
Conventional hydrodynamics describes systems with few long-lived excitations. In one dimension, however, many experimentally relevant systems feature a large number of long-lived excitations and conserved quantities even at high temperature, because they are proximate to integrable limits. Such models cannot be treated using conventional hydrodynamics. The framework of generalized hydrodynamics (GHD) was recently developed to treat the dynamics of one-dimensional models: It combines ideas from integrability, hydrodynamics, and kinetic theory to come up with a quantitative theory of transport. GHD has successfully settled several long-standing questions about one-dimensional transport; it has also been leveraged to study dynamical questions beyond the transport of conserved quantities and to systems that are not integrable. In this article, we introduce the main ideas and predictions of GHD, survey some of the most recent theoretical extensions and experimental tests of the GHD framework, and discuss some open questions in transport that the GHD perspective might elucidate
Self-Bound Superfluid Membranes and Monolayer Crystals of Ultracold Polar Molecules
We investigate the physics of ultracold dipolar molecules using path-integral quantum Monte Carlo simulations, and construct the complete phase diagram extending from weak to strong interactions and from small to mesoscopic particle numbers. Our calculations predict the formation of self-bound quantum droplets at interaction strengths lower than previous estimates for molecular condensates. Strikingly, we observe that, for stronger interactions, the oblate quantum droplet transitions to a two-dimensional sheet or superfluid membrane with a thickness of a single molecule. As interactions are increased further the system continually loses superfluidity while correlations develop, and is eventually found to undergo a transition to a crystalline monolayer that remains self-bound without external confinement. The spontaneous formation of such two-dimensional phases from a three-dimensional quantum gas is traced back to the peculiar anisotropic form of the dipole-dipole interaction generated by microwave dressing of rotational molecular states. For sufficiently large particle numbers, crystallization takes place for comparably low interaction strengths that do not promote two-body bound states and should thus be observable in ongoing experiments without limitations from three-body recombination
Mixed-initiative engineering design: towards a more-than-human design process
Product innovation is a multi-step process: a creative phase where ideas are born, an evaluation phase where the ideas are evaluated, and an implementation phase where these ideas become tangible. While computer-based assistance systems are already available for the latter two phases, creativity is often still considered an exclusively human attribute. However, recent advances in artificial intelligence (AI) have challenged this notion, as creative AI agents have demonstrated their potential to create original content. In light of these advances, a new field of research has emerged in the area of AI-enabled design processes, leading to a design process in which a computer agent collaborates with a design team to efficiently and creatively explore the entire design space in search of novel design solutions. We will present first steps towards the development of such an agent
Reinvestigation of the bound coherent neutron scattering lengths of lithium and its isotopes using neutron Bragg powder diffraction and neutron interferometry
The quality of neutron scattering studies relies upon accurately determined bound coherent neutron scattering lengthsbc, whose standard tabulated values can be in need of review. We have measured integrated Bragg peak intensities via neutron diffraction from LiF powders, as well as scattering-length densities via neutron interferometry on LiF aqueous solutions, to accurately redetermine thebcvalues for the lithium isotopes6Li and7Li, as well as that for naturally occurring elemental Li as a mix of these two isotopes. As compared to earlier studies going back to 1986, our measurements make use of considerably improved instrumentation, and careful attention was given to the experimental procedure as well as to data analysis, in particular to aspects of systematic error. From the neutron diffraction measurements we obtainbc(6Li)NPD= 2.27(2) fm,bc(7Li)NPD= -2.28(2) fm andbc(natLi)NPD= -1.94(2) fm, in excellent agreement with our interferometry results ofbc(6Li)INT= 2.28(4) fm,bc(7Li)INT= -2.28(3) fm andbc(natLi)INT= -1.93(3) fm, allowing to average them together and to propose recommended values ofbc(6Li)Rec. = 2.27(2) fm,bc(7Li)Rec. = -2.28(2) fm andbc(natLi)Rec= -1.93(2) fm, where all values correspond to pure isotopes or the standard terrestrial isotopic composition ofnatLi. Compared to standard literature values ofbc(6Li)Lit.= 2.0(1) fm,bc(7Li)Lit.= -2.22(2) fm, andbc(natLi)Lit.= -1.90(3) fm, we note in particular that our recommended value forbc(6Li) is 14% greater in magnitude, which is well outside of experimental uncertainties. Having obtained nearly identical results using two independent experimental techniques with qualitatively different possible sources of systematic error, we have confidence in our recommendedbcvalues. We discuss reasons for the observed discrepancies with respect to previously tabulated values, as well as the relative merits of the various methods to measure bound coherent neutron scattering lengths