Publications

Spectroscopy of CASSOWARY gravitationally-lensed galaxies in SDSS: characterisation of an extremely bright reionization-era analog at z = 1.42

Ramesh Mainali

Daniel P Stark

Tucker Jones

Richard S Ellis

Yashar Hezaveh

Jane R Rigby

We present new observations of sixteen bright (r = 19 − 21) gravitationally lensed galaxies at z ≃ 1 − 3 selected from the CASSOWARY s… (voir plus)urvey. Included in our sample is the z = 1.42 galaxy CSWA-141, one of the brightest known reionization-era analogs at high redshift (g=20.5), with a large sSFR (31.2 Gyr−1) and an [OIII]+Hβ equivalent width (EW[OIII] + Hβ=730 Å) that is nearly identical to the average value expected at z ≃ 7 − 8. In this paper, we investigate the rest-frame UV nebular line emission in our sample with the goal of understanding the factors that regulate strong CIII] emission. Whereas most of the sources in our sample show weak UV line emission, we find elevated CIII] in the spectrum of CSWA-141 (EWCIII]=4.6±1.9 Å) together with detections of other prominent emission lines (OIII], Si III], Fe II⋆, Mg II). We compare the rest-optical line properties of high redshift galaxies with strong and weak CIII] emission, and find that systems with the strongest UV line emission tend to have young stellar populations and nebular gas that is moderately metal-poor and highly ionized, consistent with trends seen at low and high redshift. The brightness of CSWA-141 enables detailed investigation of the extreme emission line galaxies which become common at z > 6. We find that gas traced by the CIII] doublet likely probes higher densities than that traced by [OII] and [SII]. Characterisation of the spectrally resolved Mg II emission line and several low ionization absorption lines suggests neutral gas around the young stars is likely optically thin, potentially facilitating the escape of ionizing radiation.

2023-02-07

Monthly Notices of the Royal Astronomical Society (publié)

DeltaShield: Information Theory for Human- Trafficking Detection

Catalina Vajiac

Meng-Chieh Lee

Aayushi Kulshrestha

Sacha Lévy

Namyong Park

Andreas Olligschlaeger

Cara Jones

Reihaneh Rabbany

Christos Faloutsos

2023-02-06

ACM Transactions on Knowledge Discovery from Data (publié)

A Minimax Approach Against Multi-Armed Adversarial Attacks Detection

Federica Granese

Marco Romanelli

Siddharth Garg

Pablo Piantanida

2023-02-03

ArXiv (prépublication)

AmbieGen tool at the SBST 2022 Tool Competition

Dmytro Humeniuk

Giuliano Antoniol

Foutse Khomh

AmbieGen is a tool for generating test cases for cyber-physical systems (CPS). In the context of SBST 2022 CPS tool competition, it has been… (voir plus) adapted to generating virtual roads to test a car lane keeping assist system. AmbieGen leverages a two objective NSGA-II algorithm to produce the test cases. It has achieved the highest final score, accounting for the test case efficiency, effectiveness and diversity in both testing configurations.

2023-02-02

Proceedings of the 15th Workshop on Search-Based Software Testing (publié)

Normalizing Flow Ensembles for Rich Aleatoric and Epistemic Uncertainty Modeling

Lucas Berry

David Meger

2023-02-01

ArXiv (prépublication)

A Survey on Compositional Generalization in Applications

Baihan Lin

Djallel Bouneffouf

Irina Rish

2023-02-01

ArXiv (prépublication)

Broken Neural Scaling Laws

Ethan Caballero

Kshitij Gupta

Irina Rish

David Krueger

We present a smoothly broken power law functional form (that we refer to as a Broken Neural Scaling Law (BNSL)) that accurately models &… (voir plus) extrapolates the scaling behaviors of deep neural networks (i.e. how the evaluation metric of interest varies as amount of compute used for training (or inference), number of model parameters, training dataset size, model input size, number of training steps, or upstream performance varies) for various architectures & for each of various tasks within a large & diverse set of upstream & downstream tasks, in zero-shot, prompted, & finetuned settings. This set includes large-scale vision, language, audio, video, diffusion, generative modeling, multimodal learning, contrastive learning, AI alignment, AI capabilities, robotics, out-of-distribution (OOD) generalization, continual learning, transfer learning, uncertainty estimation / calibration, OOD detection, adversarial robustness, distillation, sparsity, retrieval, quantization, pruning, fairness, molecules, computer programming/coding, math word problems, "emergent phase transitions", arithmetic, supervised learning, unsupervised/self-supervised learning, & reinforcement learning (single agent & multi-agent). When compared to other functional forms for neural scaling, this functional form yields extrapolations of scaling behavior that are considerably more accurate on this set. Moreover, this functional form accurately models & extrapolates scaling behavior that other functional forms are incapable of expressing such as the nonmonotonic transitions present in the scaling behavior of phenomena such as double descent & the delayed, sharp inflection points present in the scaling behavior of tasks such as arithmetic. Lastly, we use this functional form to glean insights about the limit of the predictability of scaling behavior. Code is available at https://github.com/ethancaballero/broken_neural_scaling_laws

2023-01-31

ICLR.cc/2023/Conference (poster)

Critic Sequential Monte Carlo

Vasileios Lioutas

Jonathan Wilder Lavington

Justice Sefas

Matthew Niedoba

Yunpeng Liu

Berend Zwartsenberg

Setareh Dabiri

Frank N. Wood

Adam Ścibior

We introduce CriticSMC, a new algorithm for planning as inference built from a composition of sequential Monte Carlo with learned Soft-Q fun… (voir plus)ction heuristic factors. These heuristic factors, obtained from parametric approximations of the marginal likelihood ahead, more effectively guide SMC towards the desired target distribution, which is particularly helpful for planning in environments with hard constraints placed sparsely in time. Compared with previous work, we modify the placement of such heuristic factors, which allows us to cheaply propose and evaluate large numbers of putative action particles, greatly increasing inference and planning efficiency. CriticSMC is compatible with informative priors, whose density function need not be known, and can be used as a model-free control algorithm. Our experiments on collision avoidance in a high-dimensional simulated driving task show that CriticSMC significantly reduces collision rates at a low computational cost while maintaining realism and diversity of driving behaviors across vehicles and environment scenarios.

2023-01-31

ICLR.cc/2023/Conference (poster)

Disentanglement of Correlated Factors via Hausdorff Factorized Support

Karsten Roth

Mark Ibrahim

Zeynep Akata

P Vincent

Diane Bouchacourt

A grand goal in deep learning research is to learn representations capable of generalizing across distribution shifts. Disentanglement is on… (voir plus)e promising direction aimed at aligning a model's representation with the underlying factors generating the data (e.g. color or background). Existing disentanglement methods, however, rely on an often unrealistic assumption: that factors are statistically independent. In reality, factors (like object color and shape) are correlated. To address this limitation, we consider the use of a relaxed disentanglement criterion -- the Hausdorff Factorized Support (HFS) criterion -- that encourages only pairwise factorized \emph{support}, rather than a factorial distribution, by minimizing a Hausdorff distance. This allows for arbitrary distributions of the factors over their support, including correlations between them. We show that the use of HFS consistently facilitates disentanglement and recovery of ground-truth factors across a variety of correlation settings and benchmarks, even under severe training correlations and correlation shifts, with in parts over

2023-01-31

ICLR.cc/2023/Conference (poster)

E3Bind: An End-to-End Equivariant Network for Protein-Ligand Docking

In silico prediction of the ligand binding pose to a given protein target is a crucial but challenging task in drug discovery. This work foc… (voir plus)uses on blind flexible selfdocking, where we aim to predict the positions, orientations and conformations of docked molecules. Traditional physics-based methods usually suffer from inaccurate scoring functions and high inference cost. Recently, data-driven methods based on deep learning techniques are attracting growing interest thanks to their efficiency during inference and promising performance. These methods usually either adopt a two-stage approach by first predicting the distances between proteins and ligands and then generating the final coordinates based on the predicted distances, or directly predicting the global roto-translation of ligands. In this paper, we take a different route. Inspired by the resounding success of AlphaFold2 for protein structure prediction, we propose E3Bind, an end-to-end equivariant network that iteratively updates the ligand pose. E3Bind models the protein-ligand interaction through careful consideration of the geometric constraints in docking and the local context of the binding site. Experiments on standard benchmark datasets demonstrate the superior performance of our end-to-end trainable model compared to traditional and recently-proposed deep learning methods.

2023-01-31

ICLR.cc/2023/Conference (poster)

Fooling SHAP with Stealthily Biased Sampling

Gabriel Laberge

Ulrich Matchi Aïvodji

Satoshi Hara

Mario Marchand

Foutse Khomh

SHAP explanations aim at identifying which features contribute the most to the difference in model prediction at a speciﬁc input versus a … (voir plus)background distribution. Recent studies have shown that they can be manipulated by malicious adversaries to produce arbitrary desired explanations. However, existing attacks focus solely on altering the black-box model itself. In this paper, we propose a complementary family of attacks that leave the model intact and manipulate SHAP explanations using stealthily biased sampling of the data points used to approximate expectations w.r.t the background distribution. In the context of fairness audit, we show that our attack can reduce the importance of a sensitive feature when explaining the difference in outcomes between groups, while remaining undetected. These results highlight the manipulability of SHAP explanations and encourage auditors to treat post-hoc explanations with skepticism.

2023-01-31

ICLR.cc/2023/Conference (poster)