Negative / Null Result ReportOpen accessComputer Science
Mykolas Sveistrys, Richard Kunert · 2025 · arXiv
Retrieval-Augmented Generation (RAG) has been used in question answering (QA) systems to improve performance when relevant information is in one (single-hop) or multiple (multi-hop) passages. However, many real life scenarios (e.g. dealing with financial, legal, medical reports) require checking all documents for relevant information without a clear stopping condition. We term these pluri-hop questions, and formalize them by 3 conditions - recall sensitivity, exhaustiveness, and exactness. To study this setting, we introduce PluriHopWIND, a multilingual diagnostic benchmark of 48 pluri-hop que
Failed Experiment ReportOpen accessEngineering
Ramin Babaee, Shahab Oveis Gharan, Martin Bouchard · 2025 · arXiv
We propose a novel digital-to-analog converter (DAC) weighting architecture that statistically minimizes the distortion caused by random timing mismatches among current sources. To decode the DAC input codewords into corresponding DAC switches, we present three algorithms with varying computational complexities. We perform high-level Matlab simulations to illustrate the dynamic performance improvement over the segmented structure.
Negative / Null Result ReportOpen accessComputer Science
Fabrizio Ottati, Giovanna Turvani, Marco Vacca et al. · 2023 · arXiv
The speed of modern digital systems is severely limited by memory latency (the ``Memory Wall'' problem). Data exchange between Logic and Memory is also responsible for a large part of the system energy consumption. Logic--In--Memory (LiM) represents an attractive solution to this problem. By performing part of the computations directly inside the memory the system speed can be improved while reducing its energy consumption. LiM solutions that offer the major boost in performance are based on the modification of the memory cell. However, what is the cost of such modifications? How do these impa
Negative / Null Result ReportOpen accessComputer Science
Yoonjeon Kim, Yuhta Takida, Chieh-Hsin Lai et al. · 2026 · arXiv
RL-based post-training has been widely adopted to enable interleaved visual and textual reasoning in unified multimodal models capable of both text and image generation. However, most existing approaches are built upon autoregressive (AR) unified models, which require full image regeneration during visual reasoning. In this work, we demonstrate that multimodal discrete diffusion models are effective alternatives to AR models for reinforcement learning in interleaved reasoning, owing to their ability to perform efficient visual rollouts via localized visual editing rather than full image-token
Negative / Null Result ReportOpen accessComputer Science
W. W. Cohen · 1995 · arXiv
In a companion paper it was shown that the class of constant-depth determinate k-ary recursive clauses is efficiently learnable. In this paper we present negative results showing that any natural generalization of this class is hard to learn in Valiant's model of pac-learnability. In particular, we show that the following program classes are cryptographically hard to learn: programs with an unbounded number of constant-depth linear recursive clauses; programs with one constant-depth determinate clause containing an unbounded number of recursive calls; and programs with one linear recursive cla
Negative / Null Result ReportOpen accessComputer Science
Tim Menzies, Ye Yang, George Mathew et al. · 2016 · arXiv
Context:More than half the literature on software effort estimation (SEE) focuses on comparisons of new estimation methods. Surprisingly, there are no studies comparing state of the art latest methods with decades-old approaches. Objective:To check if new SEE methods generated better estimates than older methods. Method: Firstly, collect effort estimation methods ranging from "classical" COCOMO (parametric estimation over a pre-determined set of attributes) to "modern" (reasoning via analogy using spectral-based clustering plus instance and feature selection, and a recent "baseline method" pro
Negative / Null Result ReportOpen accessComputer Science
Hui Chen, Deepanway Ghosal, Navonil Majumder et al. · 2020 · arXiv
Persuasion aims at forming one's opinion and action via a series of persuasive messages containing persuader's strategies. Due to its potential application in persuasive dialogue systems, the task of persuasive strategy recognition has gained much attention lately. Previous methods on user intent recognition in dialogue systems adopt recurrent neural network (RNN) or convolutional neural network (CNN) to model context in conversational history, neglecting the tactic history and intra-speaker relation. In this paper, we demonstrate the limitations of a Transformer-based approach coupled with Co
Negative / Null Result ReportOpen accessMathematics
Francesco Orabona, Ryan D'Orazio · 2025 · arXiv
The Polyak stepsize has been proven to be a fundamental stepsize in convex optimization, giving near optimal gradient descent rates across a wide range of assumptions. The universality of the Polyak stepsize has also inspired many stochastic variants, with theoretical guarantees and strong empirical performance. Despite the many theoretical results, our understanding of the convergence properties and shortcomings of the Polyak stepsize or its variants is both incomplete and fractured across different analyses. We propose a new, unified, and simple perspective for the Polyak stepsize and its va
Negative / Null Result ReportOpen accessComputer Science
Alankar Kotwal, Ajit Rajwade · 2017 · arXiv
The bound that arises out of sparse recovery analysis in compressed sensing involves input signal sparsity and some property of the sensing matrix. An effort has therefore been made in the literature to optimize sensing matrices for optimal recovery using this property. We discover, in the specific case of optimizing codes for the CACTI camera, that the popular method of mutual coherence minimization does not produce optimal results: codes designed to optimize effective dictionary coherence often perform worse than random codes in terms of mean squared reconstruction error. This surprising phe
Negative / Null Result ReportOpen accessComputer Science
Maxim Eingorn, Alexander Zhuk · 2010 · arXiv
In Kaluza-Klein model with toroidal extra dimensions, we obtain the metric coefficients in a weak-field approximation for delta-shaped matter sources. These metric coefficients are applied to calculate the formulas for frequency shift, perihelion shift, deflection of light and parameterized post-Newtonian (PPN) parameters. In the leading order of approximation, the formula for frequency shift coincides with well-known general relativity expression. However, for perihelion shift, light deflection and PPN parameter $γ$ we obtain formulas $Dπr_g/[(D-2)a(1-e^2)]$, $(D-1)r_g/[(D-2)ρ]$ and $1/(D-2)$
Negative / Null Result ReportOpen accessMathematics
Jose Moncayo, Pedro H. Zambrano · 2023 · arXiv
We investigate different set-theoretic constructions in Residuated Logic based on Fitting's work on Intuitionistic Set Theory. We start by stating some results concerning constructible sets within valued models of Set Theory. We present two distinct constructions of the constructible universe: $\mathfrak{L}^{\mathbb{Q}}$ and $\mathbb{L}^{\mathbb{Q}}$, and show that they are isomorphic to V (the classical von Neumann universe) and L (the classical Gödel constructible universe), respectively. Even though lattice-valued models are the natural way to study non-classical Set Theory (e.g., Intuition
Negative / Null Result ReportOpen accessComputer Science
Lukas Balles, Cedric Archambeau, Giovanni Zappella · 2023 · arXiv
With increasing scale in model and dataset size, the training of deep neural networks becomes a massive computational burden. One approach to speed up the training process is Selective Backprop. For this approach, we perform a forward pass to obtain a loss value for each data point in a minibatch. The backward pass is then restricted to a subset of that minibatch, prioritizing high-loss examples. We build on this approach, but seek to improve the subset selection mechanism by choosing the (weighted) subset which best matches the mean gradient over the entire minibatch. We use the gradients w.r
Negative / Null Result ReportOpen accessComputer Science
Simon Kuang, Yuezhu Xu, S. Sivaranjani et al. · 2026 · arXiv
The global Lipschitz constant of a neural network is related to robustness and generalization, yet unlike in many classical models, it is not plainly legible from the parameters. This has motivated sophisticated verification algorithms, especially semidefinite programming (SDP) based on incremental quadratic constraints on the activation functions, to improve on the fast but often loose product of layerwise Lipschitz constants (the trivial bound). We ask why Lipschitz verification is a problem in the first place. Our answer is that the difficulty is structural: estimating a network's Lipschitz
Negative / Null Result ReportOpen accessComputer Science
Kamyar Azizzadenesheli, Brandon Yang, Weitang Liu et al. · 2018 · arXiv
While many recent advances in deep reinforcement learning (RL) rely on model-free methods, model-based approaches remain an alluring prospect for their potential to exploit unsupervised data to learn environment model. In this work, we provide an extensive study on the design of deep generative models for RL environments and propose a sample efficient and robust method to learn the model of Atari environments. We deploy this model and propose generative adversarial tree search (GATS) a deep RL algorithm that learns the environment model and implements Monte Carlo tree search (MCTS) on the lear
Negative / Null Result ReportOpen accessMathematics
Christoph Schultheiss, Peter Bühlmann · 2022 · arXiv
We consider likelihood score-based methods for causal discovery in structural causal models. In particular, we focus on Gaussian scoring and analyze the effect of model misspecification in terms of non-Gaussian error distribution. We present a surprising negative result for Gaussian likelihood scoring in combination with nonparametric regression methods.
Failed Experiment ReportOpen accessComputer Science
Antoine Suarez · 2014 · arXiv
A Michelson-Morley-type experiment is described, which exploits two-photon interference between entangled photons instead of classical light interference. In this experimental context, the negative result (no shift in the detection rates) rules out David Bohm's postulate of an infinite-speed time-ordered "quantum potential", and thereby upholds the timeless standard quantum collapse.
Failed Experiment ReportOpen accessMathematics
Saharon Shelah · 2001 · arXiv
David Aspero asks on the possibility of having Forcing axiom FA_{aleph_2}(K), where K is the class of forcing notions preserving stationarity of subsets of aleph_1 and of aleph_2. We answer negatively, in fact we show the negative result for any regular lambda>aleph_1 even demanding adding no new sequence of ordinals of length<lambda.
Negative / Null Result ReportOpen accessPhysics
Valentin Voroshilov · 2014 · arXiv
A novel canonical transformation is offered as the mean for studying properties of a system of strongly correlated electrons. As an example of the utility of the transformation, it is used to demonstrate the existence of a quantum phase transition in a Hubbard model on a square lattice. An Appendix presents two cases with a negative result.
Negative / Null Result ReportOpen accessMathematics
Takahiro Hasebe · 2013 · arXiv
We prove that many of beta, beta prime, gamma, inverse gamma, Student t- and ultraspherical distributions are freely infinitely divisible, but some of them are not. The latter negative result follows from a local property of probability density functions. Moreover, we show that the Gaussian, ultraspherical and many of Student t-distributions have free divisibility indicator 1.
Negative / Null Result ReportOpen accessComputer Science
Samuel R. Bowman, Jennimaria Palomaki, Livio Baldini Soares et al. · 2020 · arXiv
Natural language inference (NLI) data has proven useful in benchmarking and, especially, as pretraining data for tasks requiring language understanding. However, the crowdsourcing protocol that was used to collect this data has known issues and was not explicitly optimized for either of these purposes, so it is likely far from ideal. We propose four alternative protocols, each aimed at improving either the ease with which annotators can produce sound training examples or the quality and diversity of those examples. Using these alternatives and a fifth baseline protocol, we collect and compare
Failed Experiment ReportOpen accessMathematics
Michael Joachim, Thomas Schick · 1999 · arXiv
The Gromov-Lawson-Rosenberg-conjecture for a group G states that a closed spin manifold M^n (n>4) with fundamental group G admits a metric with positive scalar curvature if and only if its C^*-index A(M) in KO_n(C^*_r(G)) vanishes. We prove this for groups G with low-dimensional classifying space, provided the assembly map for G is injective. On the other hand, we construct a spin manifold with no metric with scal>0 but so that already its KO-orientation in KO_*(B pi_1(M)) vanishes. Therefore a corresponding weakened version or the GLR-conjecture is wrong. Last we address non-orientable manifo
Negative / Null Result ReportOpen accessComputer Science
Raphaël Berthon, Nathanaël Fijalkow, Emmanuel Filiot et al. · 2020 · arXiv
We study alternating automata with qualitative semantics over infinite binary trees: alternation means that two opposing players construct a decoration of the input tree called a run, and the qualitative semantics says that a run of the automaton is accepting if almost all branches of the run are accepting. In this paper we prove a positive and a negative result for the emptiness problem of alternating automata with qualitative semantics. The positive result is the decidability of the emptiness problem for the case of Büchi acceptance condition. An interesting aspect of our approach is that we
Negative / Null Result ReportOpen accessMathematics
Matthew D. Kvalheim · 2026 · arXiv
We study whether second-order systems can be made to behave like prescribed first-order dynamical systems through feedback control. More precisely, we study whether prescribed vector fields on compact smooth manifolds, viewed geometrically as sections of the tangent bundle, can be asymptotically stabilized in a strong sense by second-order control systems on the base manifold. Our class of second-order systems includes most Lagrangian systems, and we obtain both positive and negative results. The positive result asserts that, for fully actuated systems, the section corresponding to any smooth
Negative / Null Result ReportOpen accessMathematics
László Csató · 2017 · arXiv
It has been shown recently that the Eigenvector Method may lead to strong rank reversal in group decision making, that is, the alternative with the highest priority according to all individual vectors may lose its position when evaluations are derived from the aggregated group comparison matrix. We give a minimal counterexample and prove that this negative result is a consequence of the difference of the rankings induced by the right and inverse left eigenvectors.
Negative / Null Result ReportOpen accessComputer Science
Maryam Aliakbarpour, Rose Silver, Thomas Steinke et al. · 2023 · arXiv
We construct differentially private estimators with low sample complexity that estimate the median of an arbitrary distribution over $\mathbb{R}$ satisfying very mild moment conditions. Our result stands in contrast to the surprising negative result of Bun et al. (FOCS 2015) that showed there is no differentially private estimator with any finite sample complexity that returns any non-trivial approximation to the median of an arbitrary distribution.
Negative / Null Result ReportOpen accessComputer Science
Hayato Takahashi · 2011 · arXiv
A new negative result for nonparametric distribution estimation of binary ergodic processes is shown. The problem of estimation of distribution with any degree of accuracy is studied. Then it is shown that for any countable class of estimators there is a zero-entropy binary ergodic process that is inconsistent with the class of estimators. Our result is different from other negative results for universal forecasting scheme of ergodic processes. We also introduce a related result by B. Weiss.
Negative / Null Result ReportOpen accessMathematics
Geivison Ribeiro · 2024 · arXiv
This note presents an extension of a result within the concept of [S]-lineability, originally developed in 2019 by L. Bernal-González, J.A. Conejero, M. Murillo-Arcila, and J.B. Seoane-Sepúlveda . Additionally, we provide a characterization in terms of lineability in the context of complements of unions of closed subspaces in F-spaces, and finally, we present a negative result in both normed spaces and p-Banach spaces. These findings contribute to the understanding of linearity in exotic settings in topological vector spaces.
Negative / Null Result ReportOpen accessComputer Science
Pierre-Étienne Meunier, Matthew J. Patitz, Scott M. Summers et al. · 2013 · arXiv
We prove a negative result on the power of a model of algorithmic self-assembly for which it has been notoriously difficult to find general techniques and results. Specifically, we prove that Winfree's abstract Tile Assembly Model, when restricted to use noncooperative tile binding, is not intrinsically universal. This stands in stark contrast to the recent result that, via cooperative binding, the abstract Tile Assembly Model is indeed intrinsically universal. Noncooperative self-assembly, also known as "temperature 1", is where tiles bind to each other if they match on one or more sides, whe
Negative / Null Result ReportOpen accessMathematics
Jack Hall, Amnon Neeman, David Rydh · 2014 · arXiv
Let $X$ be a quasi-compact and quasi-separated scheme. There are two fundamental and pervasive facts about the unbounded derived category of $X$: (1) $\mathsf{D}_{\mathrm{qc}}(X)$ is compactly generated by perfect complexes and (2) if $X$ is noetherian or has affine diagonal, then the functor $Ψ_X \colon \mathsf{D}(\mathsf{QCoh}(X)) \to \mathsf{D}_{\mathrm{qc}}(X)$ is an equivalence. Our main results are that for algebraic stacks in positive characteristic, the assertions (1) and (2) are typically false.
Negative / Null Result ReportOpen accessMathematics
Michael Edward Hill · 2021 · arXiv
In a pervious paper Weidmann shows that there a bound on the number of orbits of edges in a tree on which a finitely generated group acts $(k,C)$-acylindrically. In this paper we extend this result to actions which are $k$-acylindrical except on a family of groups with "finite height". We also give an example which gives a negative result to a conjecture of Weidmann from the same paper and produce a sharp bound for groups acting $k$--acylindrically.
