activity
20142026
most citedWhere is the Information in a Deep Neural Network?

46 citations · 160 across the 23 of their papers we have counts for

collaborators

24 papers

math.GR2026

The conjecture for Artin groups of spherical type

Giovanni Paolini

In these notes, we introduce the 50-year-old conjecture alongside Coxeter and Artin groups. Roughly speaking, the conjecture states that the complement in …

cs.CL2023★ 2 cited

Taxonomy Expansion for Named Entity Recognition

Karthikeyan K, Yogarshi Vyas, Jie Ma +7

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 addition…

cs.CL2023

A Weak Supervision Approach for Few-Shot Aspect Based Sentiment

Robert Vacareanu, Siddharth Varia, Kishaloy Halder +5

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 a…

cs.LG2023

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

Benjamin Bowman, Alessandro Achille, Luca Zancato +4

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…

cs.LG2022

Stacked Residuals of Dynamic Layers for Time Series Anomaly Detection

L. Zancato, A. Achille, G. Paolini +2

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…

math.GR2021

The dual approach to the conjecture

Giovanni Paolini

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 conjectur…

cs.LG2021

DIVA: Dataset Derivative of a Learning Task

Yonatan Dukler, Alessandro Achille, Giovanni Paolini +3

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…

math.GR2021

Factoring isometries of quadratic spaces into reflections

Jon McCammond, Giovanni Paolini

Let be a vector space endowed with a non-degenerate quadratic form . If the base field is different from , it is known that every isometry can be…

cs.LG2021★ 2 cited

Estimating informativeness of samples with Smooth Unique Information

Hrayr Harutyunyan, Alessandro Achille, Giovanni Paolini +4

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 w…

cs.LG2021★ 36 cited

Structured Prediction as Translation between Augmented Natural Languages

Giovanni Paolini, Ben Athiwaratkun, Jason Krone +6

We propose a new framework, Translation between Augmented Natural Languages (TANL), to solve many structured prediction language tasks including joint entity and relation extractio…

math.GT2020

A table of -component handlebody links of genus up to six crossings

Giovanni Bellettini, Giovanni Paolini, Maurizio Paolini +1

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 handl…

math.CO2019★ 7 cited

Representations of torsion-free arithmetic matroids

Roberto Pagaria, Giovanni Paolini

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…

math.GR2019★ 32 cited

Proof of the conjecture for affine Artin groups

Giovanni Paolini, Mario Salvetti

We prove the conjecture for affine Artin groups: the complexified complement of an affine reflection arrangement is a classifying space. This is a long-standing problem, d…

cs.LG2019★ 46 cited

Where is the Information in a Deep Neural Network?

Alessandro Achille, Giovanni Paolini, Stefano Soatto

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…

cs.LG2019

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

Alessandro Achille, Giovanni Paolini, Glen Mbeng +1

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 learni…

math.CO2018★ 3 cited

Shellability of generalized Dowling posets

Giovanni Paolini

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 po…

q-bio.PE2018★ 2 cited

Impossibility results on stability of phylogenetic consensus methods

Emanuele Delucchi, Linard Hoessly, Giovanni Paolini

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 des…

math.AT2018★ 9 cited

Euclidean matchings and minimality of hyperplane arrangements

Davide Lofano, Giovanni Paolini

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 pr…

math.AT2017★ 1 cited

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

Giovanni Paolini

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 cal…

math.CO2017

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

Emanuele Delucchi, Noriane Girard, Giovanni Paolini

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. Ou…

cs.CG2017

Collapsibility to a subcomplex of a given dimension is NP-complete

Giovanni Paolini

In this paper we extend the works of Tancer and of Malgouyres and Francés, showing that -collapsibility is NP-complete for except . By -collapsibil…

math.AT2017★ 15 cited

Weighted sheaves and homology of Artin groups

Giovanni Paolini, Mario Salvetti

In this paper we expand the theory of weighted sheaves over posets, and use it to study the local homology of Artin groups. First, we use such theory to relate the homology of clas…

math.AT2015★ 4 cited

On the classifying space of Artin monoids

Giovanni Paolini

A theorem proved by Dobrinskaya in 2006 shows that there is a strong connection between the conjecture for Artin groups and the classifying space of Artin monoids. More re…

cs.DS2014★ 1 cited

An algorithm for canonical forms of finite subsets of up to affinities

Giovanni Paolini

In this paper we describe an algorithm for the computation of canonical forms of finite subsets of , up to affinities over . For fixed dimension , this…