Daily Summary for 2026-11-11
daily
Transcript
Introduction to the show: ident: High Energy Physics: Lattice. The day in lattice field theory: what the simulations put on the table.
Kai: It's the eleventh of November, twenty twenty-six, and this is the day's research.
Mira: 2 new papers came out today.
Kai: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We'll take the day in one pass, then pull out the papers we're staying with.
Mira: Start us off with Negative-Power Density of States in Chiral-Orthogonal Disordered Systems and the IR Phase of QCD.
The day's research: Kai: We are looking at quantum systems with disorder and infrared phases in QCD.
Mira: Specifically, negative power density of states emerges in chiral Anderson models.
Lev: The work suggests power p one vanishes even with moderate disorder strength.
Kai: Instead, the spectrum shows a power behavior with p less than zero near E equals one.
Mira: This connects to random hoppings from a scale-invariant 1 over t distribution.
Lev: That distribution regularized between e minus W over two and e W over two produces negative power density of states.
Kai: Another paper focuses on computing these in lattice QCD using variance reduction techniques.
Mira: Variance reduction with probing and Multilevel Monte Carlo shows multigrid multilevel Monte Carlo is best for low modes.
Lev: Stochastic probing is superior when dominant fluctuations are short-ranged and local.
Kai: This shows complementarity between the two methods, long distance versus localized quantities.
Mira: Deflation schemes work best for observables dominated by long distance propagation.
Lev: Probing is better for localized quantities; combining them targets both contributions to variance.
Lucky paper: 2610.35524: Kai: So we’re looking at the paper on Negative-Power Density of States in Chiral-Orthogonal Disordered Systems and the IR Phase of QCD.
Mira: It investigates whether negative power law behavior in Dirac spectral density near zero can mimic the infrared phase of Quantum Chromodynamics.
Lev: The authors connect scale invariant noise to scale invariant eigenstate distributions using a cutoff W.
Kai: They use chiral orthogonal Anderson models where hopping strengths come from a distribution regularized by W, specifically the one over t distribution between e minus W over two and eW over two.
Mira: This study shows that at finite disorder strength W of moderate strength, these models don't have a power singular density of states at zero energy in the thermodynamic limit.
Lev: Instead, the spectrum shows a very accurate power behavior with p less than zero on an interval around E equals one that spreads closer to both zero and infinity as W increases.
Kai: That means the spectral features of the model Hamiltonian are imprinted by that noise regularization interval between e minus W over two and eW over two.
Mira: They show that at least in some range of finite W, the power near E equals zero vanishes in the thermodynamic limit, so there isn't a true negative power divergence in density of states.
Lev: When they remove that noise cutoff and let W go to infinity, quantum mechanics turns the scale invariant one over t noise into a scale invariant one over E distribution of eigenstates.
Kai: Which mimics the interpretation of scale invariant Dirac spectral densities in the infrared phase as being due to scale invariant glue backgrounds.
Mira: The authors conclude that modeling the IR sector requires viewing that IR cutoff not as the lattice size L but descending from the randomness strength W.
Lev: Their results strongly suggest that p E equals minus one for all E greater than zero when W goes to infinity.
Kai: They also show that the convergence to p E equals minus one is uniform on any fixed positive energy interval between two points, E1 and E2.
Lucky paper: 2607.05157: Kai: So we're moving on to Variance reduction with probing and Multilevel Monte Carlo in Lattice QCD.
Mira: This paper investigates two complementary variance reduction strategies for estimating traces in lattice QCD computations.
Lev: The results show that multigrid multilevel Monte Carlo is most effective on observables that are strongly influenced by low modes.
Kai: But what about when the fluctuations are more localized?
Mira: That's where stochastic probing comes in; it's most effective when the dominant fluctuations are short-ranged and local.
Lev: The multigrid MLMC method uses a multigrid hierarchy to build an unbiased multilevel estimator through recursive coarse grid corrections.
Kai: So, how does this relate to the actual cost reduction they saw?
Mira: For the connected pseudoscalar two-point function, the multilevel decomposition yields a variance reduction of up to ten five at large time separations.
Lev: That translates into a clear cost reduction at fixed accuracy for that specific observable.
Kai: What about disconnected loops, since we talked about those earlier?
Mira: For disconnected loops, probing combined with dilution delivers a substantial cost reduction that improves as the number of probing vectors is increased.
Lev: The observed complementarity motivates studying combined strategies, specifically integrating multigrid MLMC with the proposed probing approach to target both long-distance and short-distance contributions to variance simultaneously.
More episodes
- The day in conversation for 2026-11-19
- Daily Summary for 2026-11-19
- 2606.20029-Finite-Volume Step Scaling at Fixed Truncation in (2+1)D Hamiltonian U(1) Lattice Gauge Theory
- 2603.16266-Lattice QCD study of the K*(892) resonance at the physical point
- 2607.15742-Path optimization method for the sign problem: Insights from random matrix models
- The day in conversation for 2026-11-18
- Daily Summary for 2026-11-18
- Daily Summary for 2026-11-17
- 2608.18239-Searching for symmetric mass generation with staggered fermions in four dimensions
- The day in conversation for 2026-11-17