arXiv subjects

Francesco Tonolini

Publications and source records attributed to Francesco Tonolini.

12 recordsLinked to original sources

arXiv2025

Measuring Uncertainty Calibration

We make two contributions to the problem of estimating the $L_1$ calibration error of a binary classifier from a finite dataset. First, we provide an upper bound for any classifier where the calibration function has bounded variation. Second, we provide a method of modifying any classifier so that its calibration error can be upper bounded efficiently without significantly impacting classifier performance and without any restrictive assumptions. All our results are non-asymptotic and distribution-free. We conclude by providing advice on how to measure calibration error in practice. Our methods yield practical procedures that can be run on real-world datasets with modest overhead.

cs.LG↗

arXiv2025

Bridging the Gap: From Ad-hoc to Proactive Search in Conversations

Proactive search in conversations (PSC) aims to reduce user effort in formulating explicit queries by proactively retrieving useful relevant information given conversational context. Previous work in PSC either directly uses this context as input to off-the-shelf ad-hoc retrievers or further fine-tunes them on PSC data. However, ad-hoc retrievers are pre-trained on short and concise queries, while the PSC input is longer and noisier. This input mismatch between ad-hoc search and PSC limits retrieval quality. While fine-tuning on PSC data helps, its benefits remain constrained by this input gap. In this work, we propose Conv2Query, a novel conversation-to-query framework that adapts ad-hoc retrievers to PSC by bridging the input gap between ad-hoc search and PSC. Conv2Query maps conversational context into ad-hoc queries, which can either be used as input for off-the-shelf ad-hoc retrievers or for further fine-tuning on PSC data. Extensive experiments on two PSC datasets show that Conv2Query significantly improves ad-hoc retrievers' performance, both when used directly and after fine-tuning on PSC.

cs.IR↗

arXiv2023

Rethinking Semi-supervised Learning with Language Models

Semi-supervised learning (SSL) is a popular setting aiming to effectively utilize unlabelled data to improve model performance in downstream natural language processing (NLP) tasks. Currently, there are two popular approaches to make use of unlabelled data: Self-training (ST) and Task-adaptive pre-training (TAPT). ST uses a teacher model to assign pseudo-labels to the unlabelled data, while TAPT continues pre-training on the unlabelled data before fine-tuning. To the best of our knowledge, the effectiveness of TAPT in SSL tasks has not been systematically studied, and no previous work has directly compared TAPT and ST in terms of their ability to utilize the pool of unlabelled data. In this paper, we provide an extensive empirical study comparing five state-of-the-art ST approaches and TAPT across various NLP tasks and data sizes, including in- and out-of-domain settings. Surprisingly, we find that TAPT is a strong and more robust SSL learner, even when using just a few hundred unlabelled samples or in the presence of domain shifts, compared to more sophisticated ST approaches, and tends to bring greater improvements in SSL than in fully-supervised settings. Our further analysis demonstrates the risks of using ST approaches when the size of labelled or unlabelled data is small or when domain shifts exist. We offer a fresh perspective for future SSL research, suggesting the use of unsupervised pre-training objectives over dependency on pseudo labels.

cs.CL↗

arXiv2021

Bayesian parameter estimation using conditional variational autoencoders for gravitational-wave astronomy

Gravitational wave (GW) detection is now commonplace and as the sensitivity of the global network of GW detectors improves, we will observe $\mathcal{O}(100)$s of transient GW events per year. The current methods used to estimate their source parameters employ optimally sensitive but computationally costly Bayesian inference approaches where typical analyses have taken between 6 hours and 5 days. For binary neutron star and neutron star black hole systems prompt counterpart electromagnetic (EM) signatures are expected on timescales of 1 second -- 1 minute and the current fastest method for alerting EM follow-up observers, can provide estimates in $\mathcal{O}(1)$ minute, on a limited range of key source parameters. Here we show that a conditional variational autoencoder pre-trained on binary black hole signals can return Bayesian posterior probability estimates. The training procedure need only be performed once for a given prior parameter space and the resulting trained machine can then generate samples describing the posterior distribution $\sim 6$ orders of magnitude faster than existing techniques.

astro-ph.IM↗

arXiv2021

Forward and Inverse models in HCI:Physical simulation and deep learning for inferring 3D finger pose

We outline the role of forward and inverse modelling approaches in the design of human--computer interaction systems. Causal, forward models tend to be easier to specify and simulate, but HCI requires solutions of the inverse problem. We infer finger 3D position $(x,y,z)$ and pose (pitch and yaw) on a mobile device using capacitive sensors which can sense the finger up to 5cm above the screen. We use machine learning to develop data-driven models to infer position, pose and sensor readings, based on training data from: 1. data generated by robots, 2. data from electrostatic simulators 3. human-generated data. Machine learned emulation is used to accelerate the electrostatic simulation performance by a factor of millions. We combine a Conditional Variational Autoencoder with domain expertise/models experimentally collected data. We compare forward and inverse model approaches to direct inference of finger pose. The combination gives the most accurate reported results on inferring 3D position and pose with a capacitive sensor on a mobile device.

cs.HC↗

arXiv2020

The role of late photons in diffuse optical imaging

The ability to image through turbid media such as organic tissues, is a highly attractive prospect for biological and medical imaging. This is challenging however, due to the highly scattering properties of tissues which scramble the image information. The earliest photons that arrive at the detector are often associated with ballistic transmission, whilst the later photons are associated with complex paths due to multiple independent scattering events and are therefore typically considered to be detrimental to the final image formation process. In this work we report on the importance of these highly diffuse, "late" photons for computational time-of-flight diffuse optical imaging. In thick scattering materials, >80 transport mean free paths, we provide evidence that including late photons in the inverse retrieval enhances the image reconstruction quality. We also show that the late photons alone have sufficient information to retrieve images of a similar quality to early photon gated data. This result emphasises the importance in the strongly diffusive regime discussed here, of fully time-resolved imaging techniques.

physics.optics↗

arXiv2020

Variational Inference for Computational Imaging Inverse Problems

Machine learning methods for computational imaging require uncertainty estimation to be reliable in real settings. While Bayesian models offer a computationally tractable way of recovering uncertainty, they need large data volumes to be trained, which in imaging applications implicates prohibitively expensive collections with specific imaging instruments. This paper introduces a novel framework to train variational inference for inverse problems exploiting in combination few experimentally collected data, domain expertise and existing image data sets. In such a way, Bayesian machine learning models can solve imaging inverse problems with minimal data collection efforts. Extensive simulated experiments show the advantages of the proposed framework. The approach is then applied to two real experimental optics settings: holographic image reconstruction and imaging through highly scattering media. In both settings, state of the art reconstructions are achieved with little collection of training data.

cs.LG↗

arXiv2020

Spatial images from temporal data

Traditional paradigms for imaging rely on the use of a spatial structure, either in the detector (pixels arrays) or in the illumination (patterned light). Removal of the spatial structure in the detector or illumination, i.e., imaging with just a single-point sensor, would require solving a very strongly ill-posed inverse retrieval problem that to date has not been solved. Here, we demonstrate a data-driven approach in which full 3D information is obtained with just a single-point, single-photon avalanche diode that records the arrival time of photons reflected from a scene that is illuminated with short pulses of light. Imaging with single-point time-of-flight (temporal) data opens new routes in terms of speed, size, and functionality. As an example, we show how the training based on an optical time-of-flight camera enables a compact radio-frequency impulse radio detection and ranging transceiver to provide 3D images.

eess.IV↗

arXiv2020

Tomographic Auto-Encoder: Unsupervised Bayesian Recovery of Corrupted Data

We propose a new probabilistic method for unsupervised recovery of corrupted data. Given a large ensemble of degraded samples, our method recovers accurate posteriors of clean values, allowing the exploration of the manifold of possible reconstructed data and hence characterising the underlying uncertainty. In this setting, direct application of classical variational methods often gives rise to collapsed densities that do not adequately explore the solution space. Instead, we derive our novel reduced entropy condition approximate inference method that results in rich posteriors. We test our model in a data recovery task under the common setting of missing values and noise, demonstrating superior performance to existing variational methods for imputation and de-noising with different real data sets. We further show higher classification accuracy after imputation, proving the advantage of propagating uncertainty to downstream tasks with our model.

cs.LG↗

arXiv2020

Computational time-of-flight diffuse optical tomography

Imaging through a strongly diffusive medium remains an outstanding challenge in particular in association with applications in biological and medical imaging. Here we propose a method based on a single-photon time-of-flight camera that allows, in combination with computational processing of the spatial and full temporal photon distribution data, to image an object embedded inside a strongly diffusive medium over more than 80 transport mean free paths. The technique is contactless and requires one second acquisition times thus allowing Hz frame rate imaging. The imaging depth corresponds to several cm of human tissue and allows one to perform deep-body imaging, here demonstrated as a proof-of-principle.

physics.optics↗

arXiv2015

Tracking hidden objects with a single-photon camera

The ability to know what is hidden around a corner or behind a wall provides a crucial advantage when physically going around the obstacle is impossible or dangerous. Previous solutions to this challenge were constrained e.g. by their physical size, the requirement of reflective surfaces or long data acquisition times. These impede both the deployment of the technology outside the laboratory and the development of significant advances, such as tracking the movement of large-scale hidden objects. We demonstrate a non-line-of-sight laser ranging technology that relies upon the ability, using only a floor surface, to send and detect light that is scattered around an obstacle. A single-photon avalanche diode (SPAD) camera detects light back-scattered from a hidden object that can then be located with centimetre precision, simultaneously tracking its movement. This non-line-of-sight laser ranging system is also shown to work at human length scales, paving the way for a variety of real-life situations.

physics.optics↗

arXiv2014

Reconstructing high-dimensional two-photon entangled states via compressive sensing

Accurately establishing the state of large-scale quantum systems is an important tool in quantum information science; however, the large number of unknown parameters hinders the rapid characterisation of such states, and reconstruction procedures can become prohibitively time-consuming. Compressive sensing, a procedure for solving inverse problems by incorporating prior knowledge about the form of the solution, provides an attractive alternative to the problem of high-dimensional quantum state characterisation. Using a modified version of compressive sensing that incorporates the principles of singular value thresholding, we reconstruct the density matrix of a high-dimensional two-photon entangled system. The dimension of each photon is equal to $d= 17$, corresponding to a system of 83521 unknown real parameters. Accurate reconstruction is achieved with approximately 2500 measurements, only 3\% of the total number of unknown parameters in the state. The algorithm we develop is fast, computationally inexpensive, and applicable to a wide range of quantum states, thus demonstrating compressive sensing as an effective technique for measuring the state of large-scale quantum systems.

quant-ph↗