Showing 1–50 of 66 results for author: Paolini, G

Around first-order rigidity of Coxeter groups

Authors: Simon André, Gianluca Paolini

Abstract: By the work of Sela, for any free group $F$, the Coxeter group $W_ 3 = \mathbb{Z}/2\mathbb{Z} \ast \mathbb{Z}/2\mathbb{Z} \ast \mathbb{Z}/2\mathbb{Z}$ is elementarily equivalent to $W_3 \ast F$, and so Coxeter groups are not closed under elementary equivalence among finitely generated groups. In this paper we show that if we restrict to models which are generated by finitely many torsion elements… ▽ More By the work of Sela, for any free group $F$, the Coxeter group $W_ 3 = \mathbb{Z}/2\mathbb{Z} \ast \mathbb{Z}/2\mathbb{Z} \ast \mathbb{Z}/2\mathbb{Z}$ is elementarily equivalent to $W_3 \ast F$, and so Coxeter groups are not closed under elementary equivalence among finitely generated groups. In this paper we show that if we restrict to models which are generated by finitely many torsion elements (finitely torsion-generated), then we can recover striking rigidity results. Our main result is that if $(W, S)$ is a Coxeter system whose irreducible components are either spherical, or affine or (Gromov) hyperbolic, and $G$ is finitely torsion-generated and elementarily equivalent to $W$, then $G$ is itself a Coxeter group. This combines results of the second author et al. from [MPS22, PS23] with the following main hyperbolic result: if $W$ is a Coxeter hyperbolic group and $G$ is $\mathrm{AE}$-equivalent to $W$ and finitely torsion-generated, then $G$ belongs to a finite collection of Coxeter groups (modulo isomorphism). Furthermore, we show that there are two hyperbolic Coxeter groups $W$ and $W'$ which are non-isomorphic but $\mathrm{AE}$-equivalent. We also show that, on other hand, if we restrict to certain specific classes of Coxeter groups then we can recover the strongest possible form of first-order rigidity, which we call first-order torsion-rigidity, namely the Coxeter group $W$ is the only finitely torsion-generated model of its theory. Crucially, we show that this form of rigidity holds for the following classes of Coxeter groups: even hyperbolic Coxeter groups and free products of one-ended or finite hyperbolic Coxeter groups. We conjecture that the same kind of phenomena occur for the whole class of Coxeter groups. In this direction, we prove that if $W$ and $W'$ are even Coxeter groups which are elementarily equivalent, then they are isomorphic. △ Less

Submitted 1 July, 2024; originally announced July 2024.

MSC Class: 03C60; 20F55

Profinite rigidity of affine Coxeter groups

Authors: Gianluca Paolini, Rizos Sklinos

Abstract: We prove that the irreducible affine Coxeter groups are first-order rigid and deduce from this that they are profinitely rigid in the absolute sense. We then show that the first-order theory of any irreducible affine Coxeter group does not have a prime model. Finally, we prove that universal Coxeter groups of finite rank are homogeneous, and that the same applies to every hyperbolic (in the sense… ▽ More We prove that the irreducible affine Coxeter groups are first-order rigid and deduce from this that they are profinitely rigid in the absolute sense. We then show that the first-order theory of any irreducible affine Coxeter group does not have a prime model. Finally, we prove that universal Coxeter groups of finite rank are homogeneous, and that the same applies to every hyperbolic (in the sense of Gromov) one-ended right-angled Coxeter group. △ Less

Submitted 2 July, 2024; v1 submitted 1 July, 2024; originally announced July 2024.

MSC Class: 03C60; 20F55

Learning to Play 7 Wonders Duel Without Human Supervision

Authors: Giovanni Paolini, Lorenzo Moreschini, Francesco Veneziano, Alessandro Iraci

Abstract: This paper introduces ZeusAI, an artificial intelligence system developed to play the board game 7 Wonders Duel. Inspired by the AlphaZero reinforcement learning algorithm, ZeusAI relies on a combination of Monte Carlo Tree Search and a Transformer Neural Network to learn the game without human supervision. ZeusAI competes at the level of top human players, develops both known and novel strategies… ▽ More This paper introduces ZeusAI, an artificial intelligence system developed to play the board game 7 Wonders Duel. Inspired by the AlphaZero reinforcement learning algorithm, ZeusAI relies on a combination of Monte Carlo Tree Search and a Transformer Neural Network to learn the game without human supervision. ZeusAI competes at the level of top human players, develops both known and novel strategies, and allows us to test rule variants to improve the game's balance. This work demonstrates how AI can help in understanding and enhancing board games. △ Less

Submitted 2 June, 2024; originally announced June 2024.

General Purpose Verification for Chain of Thought Prompting

Authors: Robert Vacareanu, Anurag Pratik, Evangelia Spiliopoulou, Zheng Qi, Giovanni Paolini, Neha Anna John, Jie Ma, Yassine Benajiba, Miguel Ballesteros

Abstract: Many of the recent capabilities demonstrated by Large Language Models (LLMs) arise primarily from their ability to exploit contextual information. In this paper, we explore ways to improve reasoning capabilities of LLMs through (1) exploration of different chains of thought and (2) validation of the individual steps of the reasoning process. We propose three general principles that a model should… ▽ More Many of the recent capabilities demonstrated by Large Language Models (LLMs) arise primarily from their ability to exploit contextual information. In this paper, we explore ways to improve reasoning capabilities of LLMs through (1) exploration of different chains of thought and (2) validation of the individual steps of the reasoning process. We propose three general principles that a model should adhere to while reasoning: (i) Relevance, (ii) Mathematical Accuracy, and (iii) Logical Consistency. We apply these constraints to the reasoning steps generated by the LLM to improve the accuracy of the final generation. The constraints are applied in the form of verifiers: the model itself is asked to verify if the generated steps satisfy each constraint. To further steer the generations towards high-quality solutions, we use the perplexity of the reasoning steps as an additional verifier. We evaluate our method on 4 distinct types of reasoning tasks, spanning a total of 9 different datasets. Experiments show that our method is always better than vanilla generation, and, in 6 out of the 9 datasets, it is better than best-of N sampling which samples N reasoning chains and picks the lowest perplexity generation. △ Less

Submitted 30 April, 2024; originally announced May 2024.

Comments: 22 pages, preprint

Fewer Truncations Improve Language Modeling

Authors: Hantian Ding, Zijian Wang, Giovanni Paolini, Varun Kumar, Anoop Deoras, Dan Roth, Stefano Soatto

Abstract: In large language model training, input documents are typically concatenated together and then split into sequences of equal length to avoid padding tokens. Despite its efficiency, the concatenation approach compromises data integrity -- it inevitably breaks many documents into incomplete pieces, leading to excessive truncations that hinder the model from learning to compose logically coherent and… ▽ More In large language model training, input documents are typically concatenated together and then split into sequences of equal length to avoid padding tokens. Despite its efficiency, the concatenation approach compromises data integrity -- it inevitably breaks many documents into incomplete pieces, leading to excessive truncations that hinder the model from learning to compose logically coherent and factually consistent content that is grounded on the complete context. To address the issue, we propose Best-fit Packing, a scalable and efficient method that packs documents into training sequences through length-aware combinatorial optimization. Our method completely eliminates unnecessary truncations while retaining the same training efficiency as concatenation. Empirical results from both text and code pre-training show that our method achieves superior performance (e.g., relatively +4.7% on reading comprehension; +16.8% in context following; and +9.2% on program synthesis), and reduces closed-domain hallucination effectively by up to 58.3%. △ Less

Submitted 2 May, 2024; v1 submitted 16 April, 2024; originally announced April 2024.

Comments: ICML 2024

The Isomorphism Problem for Oligomorphic Groups with Weak Elimination of Imaginaries

Authors: Gianluca Paolini

Abstract: In [21] it was asked if equality on the reals is sharp as a lower bound for the complexity of topological isomorphism between oligomorphic groups. We prove that under the assumption of weak elimination of imaginaries this is indeed the case. Our methods are model theoretic and they also have applications on the classical problem of reconstruction of isomorphisms of permutation groups from (topolog… ▽ More In [21] it was asked if equality on the reals is sharp as a lower bound for the complexity of topological isomorphism between oligomorphic groups. We prove that under the assumption of weak elimination of imaginaries this is indeed the case. Our methods are model theoretic and they also have applications on the classical problem of reconstruction of isomorphisms of permutation groups from (topological) isomorphisms of automorphisms groups. As a concrete application, we give an explicit description of $\mathrm{Aut}(\mathrm{GL}(V))$ for any vector space $V$ of dimension $\aleph_0$ over a finite field, in affinity with the classical description for finite dimensional spaces due to Schreier and van der Waerden. △ Less

Submitted 6 April, 2024; originally announced April 2024.

MSC Class: 03E15; 20B27; 22A05; 20F28

Computable Scott sentences and the weak Whitehead problem for finitely presented groups

Authors: Gianluca Paolini

Abstract: We prove that if $A$ is a computable Hopfian finitely presented structure, then $A$ has a computable $d$-$Σ_2$ Scott sentence if and only if the weak Whitehead problem for $A$ is decidable. We use this to infer that every hyperbolic group as well as any polycyclic-by-finite group has a computable $d$-$Σ_2$ Scott sentence, thus covering two main classes of finitely presented groups. Our proof also… ▽ More We prove that if $A$ is a computable Hopfian finitely presented structure, then $A$ has a computable $d$-$Σ_2$ Scott sentence if and only if the weak Whitehead problem for $A$ is decidable. We use this to infer that every hyperbolic group as well as any polycyclic-by-finite group has a computable $d$-$Σ_2$ Scott sentence, thus covering two main classes of finitely presented groups. Our proof also implies that every weakly Hopfian finitely presented group is strongly defined by its $\exists^+$-types, a question which arose in a different context. △ Less

Submitted 27 March, 2024; v1 submitted 29 December, 2023; originally announced January 2024.

MSC Class: 03C57; 20F10

Torsion-free abelian groups are faithfully Borel complete and pure embeddability is a complete analytic quasi-order

Authors: Gianluca Paolini, Saharon Shelah

Abstract: In [9] we proved that the space of countable torsion-free abelian groups is Borel complete. In this paper we show that our construction from [9] satisfies several additional properties of interest. We deduce from this that countable torsion-free abelian groups are faithfully Borel complete, in fact, more strongly, we can $\mathfrak{L}_{ω_1, ω}$-interpret countable graphs in them. Secondly, we show… ▽ More In [9] we proved that the space of countable torsion-free abelian groups is Borel complete. In this paper we show that our construction from [9] satisfies several additional properties of interest. We deduce from this that countable torsion-free abelian groups are faithfully Borel complete, in fact, more strongly, we can $\mathfrak{L}_{ω_1, ω}$-interpret countable graphs in them. Secondly, we show that the relation of pure embeddability (equiv., elementary embeddability) among countable models of $\mathrm{Th}(\mathbb{Z}^{(ω)})$ is a complete analytic quasi-order. △ Less

Submitted 7 December, 2023; originally announced December 2023.

MSC Class: 03E15; 20K20

Elementary Properties of Free Lattices

Authors: J. B. Nation, Gianluca Paolini

Abstract: We start a systematic analysis of the first-order model theory of free lattices. Firstly, we prove that the free lattices of finite rank are not positively indistinguishable, as there is a positive $\exists \forall$-sentence true in $\mathbf F_3$ and false in $\mathbf F_4$. Secondly, we show that every model of $\mathrm{Th}(\mathbf F_n)$ admits a canonical homomorphism into the profinite-bounded c… ▽ More We start a systematic analysis of the first-order model theory of free lattices. Firstly, we prove that the free lattices of finite rank are not positively indistinguishable, as there is a positive $\exists \forall$-sentence true in $\mathbf F_3$ and false in $\mathbf F_4$. Secondly, we show that every model of $\mathrm{Th}(\mathbf F_n)$ admits a canonical homomorphism into the profinite-bounded completion $\mathbf H_n$ of $\mathbf F_n$. Thirdly, we show that $\mathbf H_n$ is isomorphic to the Dedekind-MacNeille completion of $\mathbf F_n$, and that $\mathbf H_n$ is not positively elementarily equivalent to $\mathbf F_n$, as there is a positive $\forall\exists$-sentence true in $\mathbf H_n$ and false in $\mathbf F_n$. Finally, we show that $\mathrm{DM}(\mathbf F_n)$ is a retract of $\mathrm{Id}(\mathbf F_n)$ and that for any lattice $\mathbf K$ which satisfies Whitman's condition $\mathrm{(W)}$ and which is generated by join prime elements, the three lattices $\mathbf K$, $\mathrm{DM}(\mathbf K)$, and $\mathrm{Id}(\mathbf K)$ all share the same positive universal first-order theory. △ Less

Submitted 27 March, 2024; v1 submitted 5 October, 2023; originally announced October 2023.

Comments: 13 pages

MSC Class: 03C05; 03C64; 06B05

The construction principle and non homogeneity of uncountable relatively free groups

Authors: Davide Carolillo, Gianluca Paolini

Abstract: In [11] Sklinos proved that any uncountable free group is not $\aleph_1$-homogenenous. This was later generalized by Belegradek in [1] to torsion-free residually finite relatively free groups, leaving open whether the assumption of residual finiteness was necessary. In this paper we use methods arising from the classical analysis of relatively free groups in infinitary logic to answer Belegradek's… ▽ More In [11] Sklinos proved that any uncountable free group is not $\aleph_1$-homogenenous. This was later generalized by Belegradek in [1] to torsion-free residually finite relatively free groups, leaving open whether the assumption of residual finiteness was necessary. In this paper we use methods arising from the classical analysis of relatively free groups in infinitary logic to answer Belegradek's question in the negative. Our methods are general and they also applications in varieties with torsion, for example we show that if $V$ contains a non-solvable group, then any uncountable $V$-free group is not $\aleph_1$-homogenenous. △ Less

Submitted 22 April, 2024; v1 submitted 20 July, 2023; originally announced July 2023.

MSC Class: 20E10; 03C50

Taxonomy Expansion for Named Entity Recognition

Authors: Karthikeyan K, Yogarshi Vyas, Jie Ma, Giovanni Paolini, Neha Anna John, Shuai Wang, Yassine Benajiba, Vittorio Castelli, Dan Roth, Miguel Ballesteros

Abstract: Training a Named Entity Recognition (NER) model often involves fixing a taxonomy of entity types. However, requirements evolve and we might need the NER model to recognize additional entity types. A simple approach is to re-annotate entire dataset with both existing and additional entity types and then train the model on the re-annotated dataset. However, this is an extremely laborious task. To re… ▽ More Training a Named Entity Recognition (NER) model often involves fixing a taxonomy of entity types. However, requirements evolve and we might need the NER model to recognize additional entity types. A simple approach is to re-annotate entire dataset with both existing and additional entity types and then train the model on the re-annotated dataset. However, this is an extremely laborious task. To remedy this, we propose a novel approach called Partial Label Model (PLM) that uses only partially annotated datasets. We experiment with 6 diverse datasets and show that PLM consistently performs better than most other approaches (0.5 - 2.5 F1), including in novel settings for taxonomy expansion not considered in prior work. The gap between PLM and all other approaches is especially large in settings where there is limited data available for the additional entity types (as much as 11 F1), thus suggesting a more cost effective approaches to taxonomy expansion. △ Less

Submitted 22 May, 2023; originally announced May 2023.

A Weak Supervision Approach for Few-Shot Aspect Based Sentiment

Authors: Robert Vacareanu, Siddharth Varia, Kishaloy Halder, Shuai Wang, Giovanni Paolini, Neha Anna John, Miguel Ballesteros, Smaranda Muresan

Abstract: We explore how weak supervision on abundant unlabeled data can be leveraged to improve few-shot performance in aspect-based sentiment analysis (ABSA) tasks. We propose a pipeline approach to construct a noisy ABSA dataset, and we use it to adapt a pre-trained sequence-to-sequence model to the ABSA tasks. We test the resulting model on three widely used ABSA datasets, before and after fine-tuning.… ▽ More We explore how weak supervision on abundant unlabeled data can be leveraged to improve few-shot performance in aspect-based sentiment analysis (ABSA) tasks. We propose a pipeline approach to construct a noisy ABSA dataset, and we use it to adapt a pre-trained sequence-to-sequence model to the ABSA tasks. We test the resulting model on three widely used ABSA datasets, before and after fine-tuning. Our proposed method preserves the full fine-tuning performance while showing significant improvements (15.84% absolute F1) in the few-shot learning scenario for the harder tasks. In zero-shot (i.e., without fine-tuning), our method outperforms the previous state of the art on the aspect extraction sentiment classification (AESC) task and is, additionally, capable of performing the harder aspect sentiment triplet extraction (ASTE) task. △ Less

Submitted 19 May, 2023; originally announced May 2023.

Anti-Classification Results for Rigidity Conditions in Abelian and Nilpotent Groups

Authors: Gianluca Paolini, Saharon Shelah

Abstract: Relying on the techniques and ideas from our recent paper [13], we prove several anti-classification results for various rigidity conditions in countable abelian and nilpotent groups. We prove three main theorems: (1) the rigid abelian groups are complete co-analytic in the space of countable torsion-free abelian groups ($\mathrm{TFAB}_ω$); (2) the Hopfian groups are complete co-analytic in… ▽ More Relying on the techniques and ideas from our recent paper [13], we prove several anti-classification results for various rigidity conditions in countable abelian and nilpotent groups. We prove three main theorems: (1) the rigid abelian groups are complete co-analytic in the space of countable torsion-free abelian groups ($\mathrm{TFAB}_ω$); (2) the Hopfian groups are complete co-analytic in $\mathrm{TFAB}_ω$; (3) the co-Hopfian groups are complete co-analytic in the space of countable $2$-nilpotent groups. In combination with our result from [13, S5], which shows that the endo-rigid abelian groups are complete co-analytic in $\mathrm{TFAB}_ω$, this shows that four major notions of rigidity from (abelian) group theory are as complex as possible as co-analytic problems. Further, the second and third theorem above solve two open questions of Thomas from [18], who asked this for the space of all countable groups. We leave open the question of whether the co-Hopfian mixed abelian groups are complete co-analytic in the space of countable abelian groups, but we reduce the problem to a concrete question on profinite groups, showing that if $G$ is a countable co-Hopfian abelian reduced group, then, for every prime number $p$, the torsion subgroup $\mathrm{Tor}_p(G)$ of $G$ is finite and $G$ embeds in the profinite group $ \prod_{p \in \mathbb{P}} \mathrm{Tor}_p(G)$. △ Less

Submitted 5 December, 2023; v1 submitted 7 March, 2023; originally announced March 2023.

MSC Class: 03E15; 20K20; 20K30

À-la-carte Prompt Tuning (APT): Combining Distinct Data Via Composable Prompting

Authors: Benjamin Bowman, Alessandro Achille, Luca Zancato, Matthew Trager, Pramuditha Perera, Giovanni Paolini, Stefano Soatto

Abstract: We introduce À-la-carte Prompt Tuning (APT), a transformer-based scheme to tune prompts on distinct data so that they can be arbitrarily composed at inference time. The individual prompts can be trained in isolation, possibly on different devices, at different times, and on different distributions or domains. Furthermore each prompt only contains information about the subset of data it was exposed… ▽ More We introduce À-la-carte Prompt Tuning (APT), a transformer-based scheme to tune prompts on distinct data so that they can be arbitrarily composed at inference time. The individual prompts can be trained in isolation, possibly on different devices, at different times, and on different distributions or domains. Furthermore each prompt only contains information about the subset of data it was exposed to during training. During inference, models can be assembled based on arbitrary selections of data sources, which we call "à-la-carte learning". À-la-carte learning enables constructing bespoke models specific to each user's individual access rights and preferences. We can add or remove information from the model by simply adding or removing the corresponding prompts without retraining from scratch. We demonstrate that à-la-carte built models achieve accuracy within $5\%$ of models trained on the union of the respective sources, with comparable cost in terms of training and inference time. For the continual learning benchmarks Split CIFAR-100 and CORe50, we achieve state-of-the-art performance. △ Less

Submitted 15 February, 2023; originally announced February 2023.

Comments: 13 pages, 4 figures, 8 tables

Dual structures on Coxeter and Artin groups of rank three

Authors: Emanuele Delucchi, Giovanni Paolini, Mario Salvetti

Abstract: We extend the theory of dual Coxeter and Artin groups to all rank-three Coxeter systems, beyond the previously studied spherical and affine cases. Using geometric, combinatorial, and topological techniques, we show that rank-three noncrossing partition posets are EL-shellable lattices and give rise to Garside groups isomorphic to the associated standard Artin groups. Within this framework, we prov… ▽ More We extend the theory of dual Coxeter and Artin groups to all rank-three Coxeter systems, beyond the previously studied spherical and affine cases. Using geometric, combinatorial, and topological techniques, we show that rank-three noncrossing partition posets are EL-shellable lattices and give rise to Garside groups isomorphic to the associated standard Artin groups. Within this framework, we prove the $K(π, 1)$ conjecture, the triviality of the center, and the solubility of the word problem for rank-three Artin groups. Some of our constructions apply to general Artin groups; we hope they will help develop complete solutions to the $K(π, 1)$ conjecture and other open problems in the area. △ Less

Submitted 29 June, 2022; originally announced June 2022.

Collective excitability in highly diluted networks of rotators

Authors: Gabriele Paolini, Marzena Ciszak, Francesco Marino, Simona Olmi, Alessandro Torcini

Abstract: We report on collective excitable events in a highly-diluted random network of non-excitable nodes. Excitability arises thanks to a self-sustained local adaptation mechanism that drives the system on a slow time-scale across a hysteretic phase transition involving states with different degrees of synchronization. These phenomena have been investigated for the Kuramoto model with bimodal distributi… ▽ More We report on collective excitable events in a highly-diluted random network of non-excitable nodes. Excitability arises thanks to a self-sustained local adaptation mechanism that drives the system on a slow time-scale across a hysteretic phase transition involving states with different degrees of synchronization. These phenomena have been investigated for the Kuramoto model with bimodal distribution of the natural frequencies and for the Kuramoto model with inertia and an unimodal frequency distribution. We consider global and local stimulation protocols and characterize the system response for different level of dilution. We compare the results with those obtained in the fully-coupled case showing that such collective phenomena are remarkably robust against network diluteness. △ Less

Submitted 14 June, 2022; originally announced June 2022.

Comments: 11 pages, 8 Figures

Rectangular analogues of the square paths conjecture and the univariate Delta conjecture

Authors: Alessandro Iraci, Roberto Pagaria, Giovanni Paolini, Anna Vanden Wyngaerd

Abstract: In this paper, we extend the rectangular side of the shuffle conjecture by stating a rectangular analogue of the square paths conjecture. In addition, we describe a set of combinatorial objects and one statistic that are a first step towards a rectangular extension of (the rise version of) the Delta conjecture, and of (the rise version of) the Delta square conjecture, corresponding to the case… ▽ More In this paper, we extend the rectangular side of the shuffle conjecture by stating a rectangular analogue of the square paths conjecture. In addition, we describe a set of combinatorial objects and one statistic that are a first step towards a rectangular extension of (the rise version of) the Delta conjecture, and of (the rise version of) the Delta square conjecture, corresponding to the case $q=1$ of an expected general statement. We also prove our new rectangular paths conjecture in the special case when the sides of the rectangle are coprime. △ Less

Submitted 31 May, 2022; originally announced June 2022.

Comments: 20 pages, 9 figures

MSC Class: 05E05

Journal ref: Combinatorial Theory (2023), 3(2)

Stacked Residuals of Dynamic Layers for Time Series Anomaly Detection

Authors: L. Zancato, A. Achille, G. Paolini, A. Chiuso, S. Soatto

Abstract: We present an end-to-end differentiable neural network architecture to perform anomaly detection in multivariate time series by incorporating a Sequential Probability Ratio Test on the prediction residual. The architecture is a cascade of dynamical systems designed to separate linearly predictable components of the signal such as trends and seasonality, from the non-linear ones. The former are mod… ▽ More We present an end-to-end differentiable neural network architecture to perform anomaly detection in multivariate time series by incorporating a Sequential Probability Ratio Test on the prediction residual. The architecture is a cascade of dynamical systems designed to separate linearly predictable components of the signal such as trends and seasonality, from the non-linear ones. The former are modeled by local Linear Dynamic Layers, and their residual is fed to a generic Temporal Convolutional Network that also aggregates global statistics from different time series as context for the local predictions of each one. The last layer implements the anomaly detector, which exploits the temporal structure of the prediction residuals to detect both isolated point anomalies and set-point changes. It is based on a novel application of the classic CUMSUM algorithm, adapted through the use of a variational approximation of f-divergences. The model automatically adapts to the time scales of the observed signals. It approximates a SARIMA model at the get-go, and auto-tunes to the statistics of the signal and its covariates, without the need for supervision, as more data is observed. The resulting system, which we call STRIC, outperforms both state-of-the-art robust statistical methods and deep neural network architectures on multiple anomaly detection benchmarks. △ Less

Submitted 24 February, 2022; originally announced February 2022.

The dual approach to the $K(π, 1)$ conjecture

Authors: Giovanni Paolini

Abstract: Dual presentations of Coxeter groups have recently led to breakthroughs in our understanding of affine Artin groups. In particular, they led to the proof of the $K(π, 1)$ conjecture and to the solution of the word problem. Will the "dual approach" extend to more general classes of Coxeter and Artin groups? In this paper, we describe the techniques used to prove the $K(π, 1)$ conjecture for affine… ▽ More Dual presentations of Coxeter groups have recently led to breakthroughs in our understanding of affine Artin groups. In particular, they led to the proof of the $K(π, 1)$ conjecture and to the solution of the word problem. Will the "dual approach" extend to more general classes of Coxeter and Artin groups? In this paper, we describe the techniques used to prove the $K(π, 1)$ conjecture for affine Artin groups and we ask a series of questions that are mostly open beyond the spherical and affine cases. △ Less

Submitted 9 December, 2021; originally announced December 2021.

DIVA: Dataset Derivative of a Learning Task

Authors: Yonatan Dukler, Alessandro Achille, Giovanni Paolini, Avinash Ravichandran, Marzia Polito, Stefano Soatto

Abstract: We present a method to compute the derivative of a learning task with respect to a dataset. A learning task is a function from a training set to the validation error, which can be represented by a trained deep neural network (DNN). The "dataset derivative" is a linear operator, computed around the trained model, that informs how perturbations of the weight of each training sample affect the valida… ▽ More We present a method to compute the derivative of a learning task with respect to a dataset. A learning task is a function from a training set to the validation error, which can be represented by a trained deep neural network (DNN). The "dataset derivative" is a linear operator, computed around the trained model, that informs how perturbations of the weight of each training sample affect the validation error, usually computed on a separate validation dataset. Our method, DIVA (Differentiable Validation) hinges on a closed-form differentiable expression of the leave-one-out cross-validation error around a pre-trained DNN. Such expression constitutes the dataset derivative. DIVA could be used for dataset auto-curation, for example removing samples with faulty annotations, augmenting a dataset with additional relevant samples, or rebalancing. More generally, DIVA can be used to optimize the dataset, along with the parameters of the model, as part of the training process without the need for a separate validation dataset, unlike bi-level optimization methods customary in AutoML. To illustrate the flexibility of DIVA, we report experiments on sample auto-curation tasks such as outlier rejection, dataset extension, and automatic aggregation of multi-modal data. △ Less

Submitted 18 November, 2021; originally announced November 2021.

On the Existence of Uncountable Hopfian and co-Hopfian Abelian Groups

Authors: Gianluca Paolini, Saharon Shelah

Abstract: We deal with the problem of existence of uncountable co-Hopfian abelian groups and (absolute) Hopfian abelian groups. Firstly, we prove that there are no co-Hopfian reduced abelian groups $G$ of size $< \mathfrak{p}$ with infinite $\mathrm{Tor}_p(G)$, and that in particular there are no infinite reduced abelian $p$-groups of size $< \mathfrak{p}$. Secondly, we prove that if… ▽ More We deal with the problem of existence of uncountable co-Hopfian abelian groups and (absolute) Hopfian abelian groups. Firstly, we prove that there are no co-Hopfian reduced abelian groups $G$ of size $< \mathfrak{p}$ with infinite $\mathrm{Tor}_p(G)$, and that in particular there are no infinite reduced abelian $p$-groups of size $< \mathfrak{p}$. Secondly, we prove that if $2^{\aleph_0} < λ< λ^{\aleph_0}$, and $G$ is abelian of size $λ$, then $G$ is not co-Hopfian. Finally, we prove that for every cardinal $λ$ there is a torsion-free abelian group $G$ of size $λ$ which is absolutely Hopfian, i.e., $G$ is Hopfian and $G$ remains Hopfian in every forcing extensions of the universe. △ Less

Submitted 31 August, 2023; v1 submitted 23 July, 2021; originally announced July 2021.

MSC Class: 20K30; 03E75

Factoring isometries of quadratic spaces into reflections

Authors: Jon McCammond, Giovanni Paolini

Abstract: Let $V$ be a vector space endowed with a non-degenerate quadratic form $Q$. If the base field $\mathbb{F}$ is different from $\mathbb{F}_2$, it is known that every isometry can be written as a product of reflections. In this article, we detail the structure of the poset of all minimal length reflection factorizations of an isometry. If $\mathbb{F}$ is an ordered field, we also study factorizations… ▽ More Let $V$ be a vector space endowed with a non-degenerate quadratic form $Q$. If the base field $\mathbb{F}$ is different from $\mathbb{F}_2$, it is known that every isometry can be written as a product of reflections. In this article, we detail the structure of the poset of all minimal length reflection factorizations of an isometry. If $\mathbb{F}$ is an ordered field, we also study factorizations into positive reflections, i.e., reflections defined by vectors of positive norm. We characterize such factorizations, under the hypothesis that the squares of $\mathbb{F}$ are dense in the positive elements (this includes Archimedean and Euclidean fields). In particular, we show that an isometry is a product of positive reflections if and only if its spinor norm is positive. As a final application, we explicitly describe the poset of all factorizations of isometries of the hyperbolic space. △ Less

Submitted 3 March, 2021; originally announced March 2021.

Torsion-Free Abelian Groups are Borel Complete

Authors: Gianluca Paolini, Saharon Shelah

Abstract: We prove that the Borel space of torsion-free Abelian groups with domain $ω$ is Borel complete, i.e., the isomorphism relation on this Borel space is as complicated as possible, as an isomorphism relation. This solves a long-standing open problem in descriptive set theory, which dates back to the seminal paper on Borel reducibility of Friedman and Stanley from 1989. We prove that the Borel space of torsion-free Abelian groups with domain $ω$ is Borel complete, i.e., the isomorphism relation on this Borel space is as complicated as possible, as an isomorphism relation. This solves a long-standing open problem in descriptive set theory, which dates back to the seminal paper on Borel reducibility of Friedman and Stanley from 1989. △ Less

Submitted 21 February, 2023; v1 submitted 24 February, 2021; originally announced February 2021.

MSC Class: 03E15; 20K20; 20K30

Estimating informativeness of samples with Smooth Unique Information

Authors: Hrayr Harutyunyan, Alessandro Achille, Giovanni Paolini, Orchid Majumder, Avinash Ravichandran, Rahul Bhotika, Stefano Soatto

Abstract: We define a notion of information that an individual sample provides to the training of a neural network, and we specialize it to measure both how much a sample informs the final weights and how much it informs the function computed by the weights. Though related, we show that these quantities have a qualitatively different behavior. We give efficient approximations of these quantities using a lin… ▽ More We define a notion of information that an individual sample provides to the training of a neural network, and we specialize it to measure both how much a sample informs the final weights and how much it informs the function computed by the weights. Though related, we show that these quantities have a qualitatively different behavior. We give efficient approximations of these quantities using a linearized network and demonstrate empirically that the approximation is accurate for real-world architectures, such as pre-trained ResNets. We apply these measures to several problems, such as dataset summarization, analysis of under-sampled classes, comparison of informativeness of different data sources, and detection of adversarial and corrupted examples. Our work generalizes existing frameworks but enjoys better computational properties for heavily over-parametrized models, which makes it possible to apply it to real-world networks. △ Less

Submitted 28 March, 2021; v1 submitted 17 January, 2021; originally announced January 2021.

Comments: ICLR 2021, 22 pages

Structured Prediction as Translation between Augmented Natural Languages

Authors: Giovanni Paolini, Ben Athiwaratkun, Jason Krone, Jie Ma, Alessandro Achille, Rishita Anubhai, Cicero Nogueira dos Santos, Bing Xiang, Stefano Soatto

Abstract: We propose a new framework, Translation between Augmented Natural Languages (TANL), to solve many structured prediction language tasks including joint entity and relation extraction, nested named entity recognition, relation classification, semantic role labeling, event extraction, coreference resolution, and dialogue state tracking. Instead of tackling the problem by training task-specific discri… ▽ More We propose a new framework, Translation between Augmented Natural Languages (TANL), to solve many structured prediction language tasks including joint entity and relation extraction, nested named entity recognition, relation classification, semantic role labeling, event extraction, coreference resolution, and dialogue state tracking. Instead of tackling the problem by training task-specific discriminative classifiers, we frame it as a translation task between augmented natural languages, from which the task-relevant information can be easily extracted. Our approach can match or outperform task-specific models on all tasks, and in particular, achieves new state-of-the-art results on joint entity and relation extraction (CoNLL04, ADE, NYT, and ACE2005 datasets), relation classification (FewRel and TACRED), and semantic role labeling (CoNLL-2005 and CoNLL-2012). We accomplish this while using the same architecture and hyperparameters for all tasks and even when training a single model to solve all tasks at the same time (multi-task learning). Finally, we show that our framework can also significantly improve the performance in a low-resource regime, thanks to better use of label semantics. △ Less

Submitted 2 December, 2021; v1 submitted 14 January, 2021; originally announced January 2021.

Journal ref: International Conference on Learning Representations (ICLR) 2021

Computable Scott Sentences for Quasi-Hopfian Finitely Presented Structures

Authors: Gianluca Paolini

Abstract: We prove that every quasi-Hopfian finitely presented structure $A$ has a $d$-$Σ_2$ Scott sentence, and that if in addition $A$ is computable and $Aut(A)$ satisfies a natural computable condition, then $A$ has a computable $d$-$Σ_2$ Scott sentence. This unifies several known results on Scott sentences of finitely presented structures and it is used to prove that other not previously considered alge… ▽ More We prove that every quasi-Hopfian finitely presented structure $A$ has a $d$-$Σ_2$ Scott sentence, and that if in addition $A$ is computable and $Aut(A)$ satisfies a natural computable condition, then $A$ has a computable $d$-$Σ_2$ Scott sentence. This unifies several known results on Scott sentences of finitely presented structures and it is used to prove that other not previously considered algebraic structures of interest have computable $d$-$Σ_2$ Scott sentences. In particular, we show that every right-angled Coxeter group of finite rank has a computable $d$-$Σ_2$ Scott sentence, as well as any strongly rigid Coxeter group of finite rank. Finally, we show that the free projective plane of rank $4$ has a computable $d$-$Σ_2$ Scott sentence, thus exhibiting a natural example where the assumption of quasi-Hopfianity is used (since this structure is not Hopfian). △ Less

Submitted 1 February, 2022; v1 submitted 25 October, 2020; originally announced October 2020.

Comments: 10 pages

MSC Class: 03C57; 03D45

First-Order Aspects of Coxeter Groups

Authors: Bernhard Muhlherr, Gianluca Paolini, Saharon Shelah

Abstract: We lay the foundations of the first-order model theory of Coxeter groups. Firstly, with the exception of the $2$-spherical non-affine case (which we leave open), we characterize the superstable Coxeter groups of finite rank, which we show to be essentially the Coxeter groups of affine type. Secondly, we characterize the Coxeter groups of finite rank which are domains, a central assumption in the t… ▽ More We lay the foundations of the first-order model theory of Coxeter groups. Firstly, with the exception of the $2$-spherical non-affine case (which we leave open), we characterize the superstable Coxeter groups of finite rank, which we show to be essentially the Coxeter groups of affine type. Secondly, we characterize the Coxeter groups of finite rank which are domains, a central assumption in the theory of algebraic geometry over groups, which in many respects (e.g. $λ$-stability) reduces the model theory of a given Coxeter system to the model theory of its associated irreducible components. In the second part of the paper we move to specific definability questions in right-angled Coxeter groups (RACGs) and $2$-spherical Coxeter groups. In this respect, firstly, we prove that RACGs of finite rank do not have proper elementary subgroups which are Coxeter groups, and prove further that reflection independent ones do not have proper elementary subgroups at all. Secondly, we prove that if the monoid $Sim(W, S)$ of $S$-self-similarities of $W$ is finitely generated, then $W$ is a prime model of its theory. Thirdly, we prove that in reflection independent RACGs of finite rank the Coxeter elements are type-determined. We then move to $2$-spherical Coxeter groups, proving that if $(W, S)$ is irreducible, $2$-spherical even and not affine, then $W$ is a prime model of its theory, and that if $W_Γ$ and $W_Θ$ are as in the previous sentence, then $W_Γ$ is elementary equivalent to $W_Θ$ if and only if $Γ\cong Θ$, thus solving the elementary equivalence problem for most of the $2$-spherical Coxeter groups. In the last part of the paper we focus on model theoretic applications of the notion of reflection length from Coxeter group theory, proving in particular that affine Coxeter groups are not connected. △ Less

Submitted 1 February, 2022; v1 submitted 25 October, 2020; originally announced October 2020.

Comments: 38 pages

MSC Class: 03C45; 03C68; 20F55; 51F15

A table of $n$-component handlebody links of genus $n+1$ up to six crossings

Authors: Giovanni Bellettini, Giovanni Paolini, Maurizio Paolini, Yi-Sheng Wang

Abstract: A handlebody link is a union of handlebodies of positive genus embedded in 3-space, which generalizes the notion of links in classical knot theory. In this paper, we consider handlebody links with one genus 2 handlebody and $n-1$ solid tori, $n>1$. Our main result is the complete classification of such handlebody links with six crossings or less, up to ambient isotopy. A handlebody link is a union of handlebodies of positive genus embedded in 3-space, which generalizes the notion of links in classical knot theory. In this paper, we consider handlebody links with one genus 2 handlebody and $n-1$ solid tori, $n>1$. Our main result is the complete classification of such handlebody links with six crossings or less, up to ambient isotopy. △ Less

Submitted 22 March, 2020; v1 submitted 8 March, 2020; originally announced March 2020.

Comments: 21 pages, 17 figures, 8 tables; title changed, typos corrected, added one reference

MSC Class: 57M25; 57M15; 57M27; 54C25

Representations of torsion-free arithmetic matroids

Authors: Roberto Pagaria, Giovanni Paolini

Abstract: We study the representability problem for torsion-free arithmetic matroids. By using a new operation called "reduction" and a "signed Hermite normal form", we provide and implement an algorithm to compute all the representations, up to equivalence. As an application, we disprove two conjectures about the poset of layers and the independence poset of a toric arrangement. We study the representability problem for torsion-free arithmetic matroids. By using a new operation called "reduction" and a "signed Hermite normal form", we provide and implement an algorithm to compute all the representations, up to equivalence. As an application, we disprove two conjectures about the poset of layers and the independence poset of a toric arrangement. △ Less

Submitted 12 August, 2019; originally announced August 2019.

Report number: 103272 MSC Class: 05B35; 14N20

Journal ref: European Journal of Combinatorics Volume 93, March 2021

Proof of the $K(π,1)$ conjecture for affine Artin groups

Authors: Giovanni Paolini, Mario Salvetti

Abstract: We prove the $K(π,1)$ conjecture for affine Artin groups: the complexified complement of an affine reflection arrangement is a classifying space. This is a long-standing problem, due to Arnol'd, Pham, and Thom. Our proof is based on recent advancements in the theory of dual Coxeter and Artin groups, as well as on several new results and constructions. In particular: we show that all affine noncros… ▽ More We prove the $K(π,1)$ conjecture for affine Artin groups: the complexified complement of an affine reflection arrangement is a classifying space. This is a long-standing problem, due to Arnol'd, Pham, and Thom. Our proof is based on recent advancements in the theory of dual Coxeter and Artin groups, as well as on several new results and constructions. In particular: we show that all affine noncrossing partition posets are EL-shellable; we use these posets to construct finite classifying spaces for dual affine Artin groups; we introduce new CW models for the orbit configuration spaces associated with arbitrary Coxeter groups; we construct finite classifying spaces for the braided crystallographic groups introduced by McCammond and Sulway. △ Less

Submitted 6 December, 2020; v1 submitted 26 July, 2019; originally announced July 2019.

MSC Class: 20F36; 20F55; 55P20

Journal ref: Inventiones mathematicae (2020)

Where is the Information in a Deep Neural Network?

Authors: Alessandro Achille, Giovanni Paolini, Stefano Soatto

Abstract: Whatever information a deep neural network has gleaned from training data is encoded in its weights. How this information affects the response of the network to future data remains largely an open question. Indeed, even defining and measuring information entails some subtleties, since a trained network is a deterministic map, so standard information measures can be degenerate. We measure informati… ▽ More Whatever information a deep neural network has gleaned from training data is encoded in its weights. How this information affects the response of the network to future data remains largely an open question. Indeed, even defining and measuring information entails some subtleties, since a trained network is a deterministic map, so standard information measures can be degenerate. We measure information in a neural network via the optimal trade-off between accuracy of the response and complexity of the weights, measured by their coding length. Depending on the choice of code, the definition can reduce to standard measures such as Shannon Mutual Information and Fisher Information. However, the more general definition allows us to relate information to generalization and invariance, through a novel notion of effective information in the activations of a deep network. We establish a novel relation between the information in the weights and the effective information in the activations, and use this result to show that models with low (information) complexity not only generalize better, but are bound to learn invariant representations of future inputs. These relations hinge not only on the architecture of the model, but also on how it is trained, highlighting the complex inter-dependency between the class of functions implemented by deep neural networks, the loss function used for training them from finite data, and the inductive bias implicit in the optimization. △ Less

Submitted 21 June, 2020; v1 submitted 29 May, 2019; originally announced May 2019.

Report number: UCLA-TR:190005

The Information Complexity of Learning Tasks, their Structure and their Distance

Authors: Alessandro Achille, Giovanni Paolini, Glen Mbeng, Stefano Soatto

Abstract: We introduce an asymmetric distance in the space of learning tasks, and a framework to compute their complexity. These concepts are foundational for the practice of transfer learning, whereby a parametric model is pre-trained for a task, and then fine-tuned for another. The framework we develop is non-asymptotic, captures the finite nature of the training dataset, and allows distinguishing learnin… ▽ More We introduce an asymmetric distance in the space of learning tasks, and a framework to compute their complexity. These concepts are foundational for the practice of transfer learning, whereby a parametric model is pre-trained for a task, and then fine-tuned for another. The framework we develop is non-asymptotic, captures the finite nature of the training dataset, and allows distinguishing learning from memorization. It encompasses, as special cases, classical notions from Kolmogorov complexity, Shannon, and Fisher Information. However, unlike some of those frameworks, it can be applied to large-scale models and real-world datasets. Our framework is the first to measure complexity in a way that accounts for the effect of the optimization scheme, which is critical in Deep Learning. △ Less

Submitted 14 July, 2020; v1 submitted 5 April, 2019; originally announced April 2019.

Report number: UCLA CSD180003

Strongly Minimal Steiner Systems I: Existence

Authors: John Baldwin, Gianluca Paolini

Abstract: A linear space is a system of points and lines such that any two distinct points determine a unique line; a Steiner $k$-system (for $k \geq 2$) is a linear space such that each line has size exactly $k$. Clearly, as a two-sorted structure, no linear space can be strongly minimal. We formulate linear spaces in a (bi-interpretable) vocabulary $τ$ with a single ternary relation $R$. We prove that for… ▽ More A linear space is a system of points and lines such that any two distinct points determine a unique line; a Steiner $k$-system (for $k \geq 2$) is a linear space such that each line has size exactly $k$. Clearly, as a two-sorted structure, no linear space can be strongly minimal. We formulate linear spaces in a (bi-interpretable) vocabulary $τ$ with a single ternary relation $R$. We prove that for every integer $k$ there exist $2^{\aleph_0}$-many integer valued functions $μ$ such that each $μ$ determines a distinct strongly minimal Steiner $k$-system $\mathcal{G}_μ$, whose algebraic closure geometry has all the properties of the ab initio Hrushovski construction. Thus each is a counterexample to the Zilber Trichotomy Conjecture. △ Less

Submitted 21 January, 2020; v1 submitted 8 March, 2019; originally announced March 2019.

MSC Class: 03C45; 51E10

Shellability of generalized Dowling posets

Authors: Giovanni Paolini

Abstract: A generalization of Dowling lattices was recently introduced by Bibby and Gadish, in a work on orbit configuration spaces. The authors left open the question as to whether these posets are shellable. In this paper we prove EL-shellability and use it to determine the homotopy type. Our result generalizes shellability of Dowling lattices and of posets of layers of abelian arrangements defined by roo… ▽ More A generalization of Dowling lattices was recently introduced by Bibby and Gadish, in a work on orbit configuration spaces. The authors left open the question as to whether these posets are shellable. In this paper we prove EL-shellability and use it to determine the homotopy type. Our result generalizes shellability of Dowling lattices and of posets of layers of abelian arrangements defined by root systems. We also show that subposets corresponding to invariant subarrangements are not shellable in general. △ Less

Submitted 3 December, 2023; v1 submitted 20 November, 2018; originally announced November 2018.

Journal ref: Journal of Combinatorial Theory, Series A 171, 2020

First-Order Model Theory of Free Projective Planes

Authors: Tapani Hyttinen, Gianluca Paolini

Abstract: We prove that the theory of open projective planes is complete and strictly stable, and infer from this that Marshall Hall's free projective planes $(π^n : 4 \leq n \leq ω)$ are all elementary equivalent and that their common theory is strictly stable and decidable, being in fact the theory of open projective planes. We further characterize the elementary substructure relation in the class of open… ▽ More We prove that the theory of open projective planes is complete and strictly stable, and infer from this that Marshall Hall's free projective planes $(π^n : 4 \leq n \leq ω)$ are all elementary equivalent and that their common theory is strictly stable and decidable, being in fact the theory of open projective planes. We further characterize the elementary substructure relation in the class of open projective planes, and show in particular that $(π^n : 4 \leq n \leq ω)$ is an elementary chain. We then prove that the theory of open projective planes does not have a prime model, that it has elimination of quantifiers down to Boolean combinations of existential formulas, and that it is not model complete. Finally, we characterize the forking independence relation in models of the theory and prove that the $π^n$'s ($4 \leq n \leq ω)$ are strongly type-homogeneous. △ Less

Submitted 10 February, 2020; v1 submitted 8 November, 2018; originally announced November 2018.

Comments: New version, several open questions have been solved

MSC Class: 03C45; 51E15

Some Results on Polish Groups

Authors: Gianluca Paolini, Saharon Shelah

Abstract: We prove that no quantifier-free formula in the language of group theory can define the $\aleph_1$-half graph in a Polish group, thus generalising some results from [6]. We then pose some questions on the space of groups of automorphisms of a given Borel complete class, and observe that this space must contain at least one uncountable group. Finally, we prove some results on the structure of the g… ▽ More We prove that no quantifier-free formula in the language of group theory can define the $\aleph_1$-half graph in a Polish group, thus generalising some results from [6]. We then pose some questions on the space of groups of automorphisms of a given Borel complete class, and observe that this space must contain at least one uncountable group. Finally, we prove some results on the structure of the group of automorphisms of a locally finite group: firstly, we prove that it is not the case that every group of automorphisms of a graph of power $λ$ is the group of automorphism of a locally finite group of power $λ$; secondly, we conjecture that the group of automorphisms of a locally finite group of power $λ$ has a locally finite subgroup of power $λ$, and reduce the problem to a problem on $p$-groups, thus settling the conjecture in the case $λ= \aleph_0$. △ Less

Submitted 10 November, 2019; v1 submitted 30 October, 2018; originally announced October 2018.

MSC Class: 03E15; 20K30; 20B27

Impossibility results on stability of phylogenetic consensus methods

Authors: Emanuele Delucchi, Linard Hoessly, Giovanni Paolini

Abstract: We answer two questions raised by Bryant, Francis and Steel in their work on consensus methods in phylogenetics. Consensus methods apply to every practical instance where it is desired to aggregate a set of given phylogenetic trees (say, gene evolution trees) into a resulting, "consensus" tree (say, a species tree). Various stability criteria have been explored in this context, seeking to model de… ▽ More We answer two questions raised by Bryant, Francis and Steel in their work on consensus methods in phylogenetics. Consensus methods apply to every practical instance where it is desired to aggregate a set of given phylogenetic trees (say, gene evolution trees) into a resulting, "consensus" tree (say, a species tree). Various stability criteria have been explored in this context, seeking to model desirable consistency properties of consensus methods as the experimental data are updated (e.g., more taxa, or more trees, are mapped). However, such stability conditions can be incompatible with some basic regularity properties that are widely accepted to be essential in any meaningful consensus method. Here, we prove that such an incompatibility does arise in the case of extension stability on binary trees and in the case of associative stability. Our methods combine general theoretical considerations with the use of computer programs tailored to the given stability requirements. △ Less

Submitted 13 November, 2023; v1 submitted 8 October, 2018; originally announced October 2018.

Journal ref: Systematic Biology 69 (3), pp. 557-565, 2020

On a Cardinal Invariant Related to the Haar Measure Problem

Authors: Gianluca Paolini, Saharon Shelah

Abstract: In [6], given a metrizable profinite group $G$, a cardinal invariant of the continuum $\mathfrak{fm}(G)$ was introduced, and a positive solution to the Haar Measure Problem for $G$ was given under the assumption that $\mathrm{non}(\mathcal{N}) \leq \mathfrak{fm}(G)$. We prove here that it is consistent with ZFC that there is a metrizable profinite group $G_*$ such that… ▽ More In [6], given a metrizable profinite group $G$, a cardinal invariant of the continuum $\mathfrak{fm}(G)$ was introduced, and a positive solution to the Haar Measure Problem for $G$ was given under the assumption that $\mathrm{non}(\mathcal{N}) \leq \mathfrak{fm}(G)$. We prove here that it is consistent with ZFC that there is a metrizable profinite group $G_*$ such that $\mathrm{non}(\mathcal{N}) > \mathfrak{fm}(G_*)$, thus demonstrating that the strategy of [6] does not suffice for a general solution to the Haar Measure Problem. △ Less

Submitted 20 June, 2019; v1 submitted 27 September, 2018; originally announced September 2018.

MSC Class: 03E17; 22C05; 20E18; 20A15

Euclidean matchings and minimality of hyperplane arrangements

Authors: Davide Lofano, Giovanni Paolini

Abstract: We construct a new class of maximal acyclic matchings on the Salvetti complex of a locally finite hyperplane arrangement. Using discrete Morse theory, we then obtain an explicit proof of the minimality of the complement. Our construction provides interesting insights also in the well-studied case of finite arrangements, and gives a nice geometric description of the Betti numbers of the complement.… ▽ More We construct a new class of maximal acyclic matchings on the Salvetti complex of a locally finite hyperplane arrangement. Using discrete Morse theory, we then obtain an explicit proof of the minimality of the complement. Our construction provides interesting insights also in the well-studied case of finite arrangements, and gives a nice geometric description of the Betti numbers of the complement. In particular, we solve a conjecture of Drton and Klivans on the characteristic polynomial of finite reflection arrangements. The minimal complex is compatible with restrictions, and this allows us to prove the isomorphism of Brieskorn's Lemma by a simple bijection of the critical cells. Finally, in the case of line arrangements, we describe the algebraic Morse complex which computes the homology with coefficients in an abelian local system. △ Less

Submitted 11 December, 2020; v1 submitted 7 September, 2018; originally announced September 2018.

Journal ref: Discrete Mathematics 344 (3), 2021

Invariant universality for projective planes

Authors: Filippo Calderoni, Gianluca Paolini

Abstract: We continue the work of [1, 2, 3] by analyzing the equivalence relation of bi-embeddability on various classes of countable planes, most notably the class of countable non-Desarguesian projective planes. We use constructions of the second author to show that these equivalence relations are invariantly universal, in the sense of [3], and thus in particular complete analytic. We also introduce a new… ▽ More We continue the work of [1, 2, 3] by analyzing the equivalence relation of bi-embeddability on various classes of countable planes, most notably the class of countable non-Desarguesian projective planes. We use constructions of the second author to show that these equivalence relations are invariantly universal, in the sense of [3], and thus in particular complete analytic. We also introduce a new kind of Borel reducibility relation for standard Borel G-spaces, which requires the preservation of stabilizers, and explain its connection with the notion of full embeddings commonly considered in category theory. △ Less

Submitted 14 October, 2020; v1 submitted 30 January, 2018; originally announced January 2018.

Comments: Unpublished notes, 10 pages

Automorphism Groups of Countable Stable Structures

Authors: Gianluca Paolini, Saharon Shelah

Abstract: For every countable structure $M$ we construct an $\aleph_0$-stable countable structure $N$ such that $Aut(M)$ and $Aut(N)$ are topologically isomorphic. This shows that it is impossible to detect any form of stability of a countable structure $M$ from the topological properties of the Polish group $Aut(M)$. For every countable structure $M$ we construct an $\aleph_0$-stable countable structure $N$ such that $Aut(M)$ and $Aut(N)$ are topologically isomorphic. This shows that it is impossible to detect any form of stability of a countable structure $M$ from the topological properties of the Polish group $Aut(M)$. △ Less

Submitted 19 November, 2018; v1 submitted 7 December, 2017; originally announced December 2017.

MSC Class: 03C45; 03E15; 22F50

The Class of Countable Projective Planes is Borel Complete

Authors: Gianluca Paolini

Abstract: We observe that Hall's free projective extension $P \mapsto F(P)$ of partial planes is a Borel map, and use a modification of the construction introduced in [9] to conclude that the class of countable non-Desarguesian projective planes is Borel complete. In the process, we also rediscover the main result of [7] on the realizability of every group as the group of collineations of some projective pl… ▽ More We observe that Hall's free projective extension $P \mapsto F(P)$ of partial planes is a Borel map, and use a modification of the construction introduced in [9] to conclude that the class of countable non-Desarguesian projective planes is Borel complete. In the process, we also rediscover the main result of [7] on the realizability of every group as the group of collineations of some projective plane. Finally, we use classical results of projective geometry to prove that the class of countable Pappian projective planes is Borel complete. △ Less

Submitted 14 November, 2018; v1 submitted 16 November, 2017; originally announced November 2017.

Comments: This paper has been merged with arXiv:1707.00294

MSC Class: 51A35; 03E15; 54H05; 05B35; 22F50

Polish Topologies for Graph Products of Groups

Authors: Gianluca Paolini, Saharon Shelah

Abstract: We give strong necessary conditions on the admissibility of a Polish group topology for an arbitrary graph product of groups $G(Γ, G_a)$, and use them to give a characterization modulo a finite set of nodes. As a corollary, we give a complete characterization in case all the factor groups $G_a$ are countable. We give strong necessary conditions on the admissibility of a Polish group topology for an arbitrary graph product of groups $G(Γ, G_a)$, and use them to give a characterization modulo a finite set of nodes. As a corollary, we give a complete characterization in case all the factor groups $G_a$ are countable. △ Less

Submitted 25 September, 2018; v1 submitted 16 November, 2017; originally announced November 2017.

MSC Class: 03E15; 20F65; 20B27

A New $ω$-Stable Plane

Authors: Gianluca Paolini

Abstract: We use a variation on Mason's $α$-function as a pre-dimension function to construct a not one-based $ω$-stable plane $P$ (i.e. a simple rank $3$ matroid) which does not admit an algebraic representation (in the sense of matroid theory) over any field. Furthermore, we characterize forking in $Th(P)$, we prove that algebraic closure and intrinsic closure coincide in $Th(P)$, and we show that… ▽ More We use a variation on Mason's $α$-function as a pre-dimension function to construct a not one-based $ω$-stable plane $P$ (i.e. a simple rank $3$ matroid) which does not admit an algebraic representation (in the sense of matroid theory) over any field. Furthermore, we characterize forking in $Th(P)$, we prove that algebraic closure and intrinsic closure coincide in $Th(P)$, and we show that $Th(P)$ fails weak elimination of imaginaries, and has Morley rank $ω$. △ Less

Submitted 10 January, 2020; v1 submitted 20 September, 2017; originally announced September 2017.

MSC Class: 03C45; 05B35

On the local homology of Artin groups of finite and affine type

Authors: Giovanni Paolini

Abstract: We study the local homology of Artin groups using weighted discrete Morse theory. In all finite and affine cases, we are able to construct Morse matchings of a special type (we call them "precise matchings"). The existence of precise matchings implies that the homology has a square-free torsion. This property was known for Artin groups of finite type, but not in general for Artin groups of affine… ▽ More We study the local homology of Artin groups using weighted discrete Morse theory. In all finite and affine cases, we are able to construct Morse matchings of a special type (we call them "precise matchings"). The existence of precise matchings implies that the homology has a square-free torsion. This property was known for Artin groups of finite type, but not in general for Artin groups of affine type. We also use the constructed matchings to compute the local homology in all exceptional cases, correcting some results in the literature. △ Less

Submitted 30 June, 2018; v1 submitted 5 September, 2017; originally announced September 2017.

Journal ref: Algebr. Geom. Topol. 19 (2019) 3615-3639

A Universal Homogeneous Simple Matroid of Rank $3$

Authors: Gianluca Paolini

Abstract: We construct a $\wedge$-homogeneous universal simple matroid of rank $3$, i.e. a countable simple rank~$3$ matroid $M_*$ which $\wedge$-embeds every finite simple rank $3$ matroid, and such that every isomorphism between finite $\wedge$-subgeometries of $M_*$ extends to an automorphism of $M_*$. We also construct a $\wedge$-homogeneous matroid $M_*(P)$ which is universal for the class of finite si… ▽ More We construct a $\wedge$-homogeneous universal simple matroid of rank $3$, i.e. a countable simple rank~$3$ matroid $M_*$ which $\wedge$-embeds every finite simple rank $3$ matroid, and such that every isomorphism between finite $\wedge$-subgeometries of $M_*$ extends to an automorphism of $M_*$. We also construct a $\wedge$-homogeneous matroid $M_*(P)$ which is universal for the class of finite simple rank $3$ matroids omitting a given finite projective plane $P$. We then prove that these structures are not $\aleph_0$-categorical, they have the independence property, they admit a stationary independence relation, and that their automorphism group embeds the symmetric group $Sym(ω)$. Finally, we use the free projective extension $F(M_*)$ of $M_*$ to conclude the existence of a countable projective plane embedding all the finite simple matroids of rank $3$ and whose automorphism group contains $Sym(ω)$, in fact we show that $Aut(F(M_*)) \cong Aut(M_*)$. △ Less

Submitted 3 October, 2018; v1 submitted 17 July, 2017; originally announced July 2017.

MSC Class: 03E15; 54H05; 05B35; 22F50

The Class of Non-Desarguesian Projective Planes is Borel Complete

Authors: Gianluca Paolini

Abstract: For every infinite graph $Γ$ we construct a non-Desarguesian projective plane $P^*_Γ$ of the same size as $Γ$ such that $Aut(Γ) \cong Aut(P^*_Γ)$ and $Γ_1 \cong Γ_2$ iff $P^*_{Γ_1} \cong P^*_{Γ_2}$. Furthermore, restricted to structures with domain $ω$, the map $Γ\mapsto P^*_Γ$ is Borel. On one side, this shows that the class of countable non-Desarguesian projective planes is Borel complete, and t… ▽ More For every infinite graph $Γ$ we construct a non-Desarguesian projective plane $P^*_Γ$ of the same size as $Γ$ such that $Aut(Γ) \cong Aut(P^*_Γ)$ and $Γ_1 \cong Γ_2$ iff $P^*_{Γ_1} \cong P^*_{Γ_2}$. Furthermore, restricted to structures with domain $ω$, the map $Γ\mapsto P^*_Γ$ is Borel. On one side, this shows that the class of countable non-Desarguesian projective planes is Borel complete, and thus not admitting a Ulm type system of invariants. On the other side, we rediscover the main result of [15] on the realizability of every group as the group of collineations of some projective plane. Finally, we use classical results of projective geometry to prove that the class of countable Pappian projective planes is Borel complete. △ Less

Submitted 13 September, 2018; v1 submitted 2 July, 2017; originally announced July 2017.

MSC Class: 51A35; 03E15; 54H05; 05B35; 22F50

Shellability of posets of labeled partitions and arrangements defined by root systems

Authors: Emanuele Delucchi, Noriane Girard, Giovanni Paolini

Abstract: We prove that the posets of connected components of intersections of toric and elliptic arrangements defined by root systems are EL-shellable and we compute their homotopy type. Our method rests on Bibby's description of such posets by means of "labeled partitions": after giving an EL-labeling and counting homology chains for general posets of labeled partitions, we obtain the stated results by co… ▽ More We prove that the posets of connected components of intersections of toric and elliptic arrangements defined by root systems are EL-shellable and we compute their homotopy type. Our method rests on Bibby's description of such posets by means of "labeled partitions": after giving an EL-labeling and counting homology chains for general posets of labeled partitions, we obtain the stated results by considering the appropriate subposets. △ Less

Submitted 20 June, 2017; originally announced June 2017.

Group Metrics for Graph Products of Cyclic Groups

Authors: Gianluca Paolini, Saharon Shelah

Abstract: We complement the characterization of the graph products of cyclic groups $G(Γ, \mathfrak{p})$ admitting a Polish group topology of [9] with the following result. Let $G = G(Γ, \mathfrak{p})$, then the following are equivalent: (i) there is a metric on $Γ$ which induces a separable topology in which $E_Γ$ is closed; (ii) $G(Γ, \mathfrak{p})$ is embeddable into a Polish group; (iii)… ▽ More We complement the characterization of the graph products of cyclic groups $G(Γ, \mathfrak{p})$ admitting a Polish group topology of [9] with the following result. Let $G = G(Γ, \mathfrak{p})$, then the following are equivalent: (i) there is a metric on $Γ$ which induces a separable topology in which $E_Γ$ is closed; (ii) $G(Γ, \mathfrak{p})$ is embeddable into a Polish group; (iii) $G(Γ, \mathfrak{p})$ is embeddable into a non-Archimedean Polish group. We also construct left-invariant separable group ultrametrics for $G = G(Γ, \mathfrak{p})$ and $Γ$ a closed graph on the Baire space, which is of independent interest. △ Less

Submitted 20 September, 2017; v1 submitted 7 May, 2017; originally announced May 2017.

MSC Class: 03E15; 20F65; 20B27

Polish Topologies for Graph Products of Cyclic Groups

Authors: Gianluca Paolini, Saharon Shelah

Abstract: We give a complete characterization of the graph products of cyclic groups admitting a Polish group topology, and show that they are all realizable as the group of automorphisms of a countable structure. In particular, we characterize the right-angled Coxeter groups (resp. Artin groups) admitting a Polish group topology. This generalizes results from [5], [7] and [4]. We give a complete characterization of the graph products of cyclic groups admitting a Polish group topology, and show that they are all realizable as the group of automorphisms of a countable structure. In particular, we characterize the right-angled Coxeter groups (resp. Artin groups) admitting a Polish group topology. This generalizes results from [5], [7] and [4]. △ Less

Submitted 8 January, 2018; v1 submitted 4 May, 2017; originally announced May 2017.

MSC Class: 03E15; 20F65; 20B27