Congratulations to the #ICLR2024 test of time and outstanding paper award winners

The Twelfth International Conference on Learning Representations (ICLR) is taking place this week in Vienna, Austria. During the opening of the conference, the outstanding paper award winners, and honourable mentions, were announced. The conference organisers also introduced a new award for this year: the test of time award. This award honours a paper from 2013/2014 that the programme chairs judge to have had a lasting impact.

Test of time award

Auto-Encoding Variational Bayes
Diederik Kingma, Max Welling

Abstract: How can we perform efficient inference and learning in directed probabilistic models, in the presence of continuous latent variables with intractable posterior distributions, and large datasets? We introduce a stochastic variational inference and learning algorithm that scales to large datasets and, under some mild differentiability conditions, even works in the intractable case. Our contributions are two-fold. First, we show that a reparameterization of the variational lower bound yields a lower bound estimator that can be straightforwardly optimized using standard stochastic gradient methods. Second, we show that for i.i.d. datasets with continuous latent variables per datapoint, posterior inference can be made especially efficient by fitting an approximate inference model (also called a recognition model) to the intractable posterior using the proposed lower bound estimator. Theoretical advantages are reflected in experimental results.

The committee write that: “This paper spearheaded the integration of deep learning with scalable probabilistic inference (amortized mean-field variational inference via a so-called reparameterization trick), giving rise to the Variational Autoencoder (VAE). The lasting value of this work is rooted in its elegance.”

Test of time runner up

Intriguing properties of neural networks
Christian Szegedy, Wojciech Zaremba, Ilya Sutskever, Joan Bruna, Dumitru Erhan, Ian Goodfellow, Rob Fergus

Abstract: Deep neural networks are highly expressive models that have recently achieved state of the art performance on speech and visual recognition tasks. While their expressiveness is the reason they succeed, it also causes them to learn uninterpretable solutions that could have counter-intuitive properties. In this paper we report two such properties. First, we find that there is no distinction between individual high level units and random linear combinations of high level units, according to various methods of unit analysis. It suggests that it is the space, rather than the individual units, that contains of the semantic information in the high layers of neural networks. Second, we find that deep neural networks learn input-output mappings that are fairly discontinuous to a significant extend. We can cause the network to misclassify an image by applying a certain imperceptible perturbation, which is found by maximizing the network’s prediction error. In addition, the specific nature of these perturbations is not a random artifact of learning: the same perturbation can cause a different network, that was trained on a different subset of the dataset, to misclassify the same input.

This paper was recognised because it: “highlighted the issue that neural networks can be vulnerable to small almost imperceptible variations to the input. This idea helped spawn the area of adversarial attacks (trying to fool a neural network) as well as adversarial defense (training a neural network to not be fooled).”

Outstanding papers

Generalization in diffusion models arises from geometry-adaptive harmonic representations
Zahra Kadkhodaie, Florentin Guth, Eero P Simoncelli, Stéphane Mallat

Abstract: Deep neural networks (DNNs) trained for image denoising are able to generate high-quality samples with score-based reverse diffusion algorithms. These impressive capabilities seem to imply an escape from the curse of dimensionality, but recent reports of memorization of the training set raise the question of whether these networks are learning the “true” continuous density of the data. Here, we show that two DNNs trained on non-overlapping subsets of a dataset learn nearly the same score function, and thus the same density, when the number of training images is large enough. In this regime of strong generalization, diffusion-generated images are distinct from the training set, and are of high visual quality, suggesting that the inductive biases of the DNNs are well-aligned with the data density. We analyze the learned denoising functions and show that the inductive biases give rise to a shrinkage operation in a basis adapted to the underlying image. Examination of these bases reveals oscillating harmonic structures along contours and in homogeneous regions. We demonstrate that trained denoisers are inductively biased towards these geometry-adaptive harmonic bases since they arise not only when the network is trained on photographic images, but also when it is trained on image classes supported on low-dimensional manifolds for which the harmonic basis is suboptimal. Finally, we show that when trained on regular image classes for which the optimal basis is known to be geometry-adaptive and harmonic, the denoising performance of the networks is near-optimal.

Learning Interactive Real-World Simulators
Sherry Yang, Yilun Du, Seyed Kamyar Seyed Ghasemipour, Jonathan Tompson, Leslie Pack Kaelbling, Dale Schuurmans, Pieter Abbeel

Abstract: Generative models trained on internet data have revolutionized how text, image, and video content can be created. Perhaps the next milestone for generative models is to simulate realistic experience in response to actions taken by humans, robots, and other interactive agents. Applications of a real-world simulator range from controllable content creation in games and movies, to training embodied agents purely in simulation that can be directly deployed in the real world. We explore the possibility of learning a universal simulator (UniSim) of real-world interaction through generative modeling. We first make the important observation that natural datasets available for learning a real-world simulator are often rich along different axes (e.g., abundant objects in image data, densely sampled actions in robotics data, and diverse movements in navigation data). With careful orchestration of diverse datasets, each providing a different aspect of the overall experience, UniSim can emulate how humans and agents interact with the world by simulating the visual outcome of both high-level instructions such as “open the drawer” and low-level controls such as “move by x,y” from otherwise static scenes and objects. There are numerous use cases for such a real-world simulator. As an example, we use UniSim to train both high-level vision-language planners and low-level reinforcement learning policies, each of which exhibit zero-shot real-world transfer after training purely in a learned real-world simulator. We also show that other types of intelligence such as video captioning models can benefit from training with simulated experience in UniSim, opening up even wider applications.

Never Train from Scratch: Fair Comparison of Long-Sequence Models Requires Data-Driven Priors
Ido Amos, Jonathan Berant, Ankit Gupta

Abstract: Modeling long-range dependencies across sequences is a longstanding goal in machine learning and has led to architectures, such as state space models, that dramatically outperform Transformers on long sequences. However, these impressive empirical gains have been by and large demonstrated on benchmarks (e.g. Long Range Arena), where models are randomly initialized and trained to predict a target label from an input sequence. In this work, we show that random initialization leads to gross overestimation of the differences between architectures and that pretraining with standard denoising objectives, using only the downstream task data, leads to dramatic gains across multiple architectures and to very small gaps between Transformers and state space models (SSMs). In stark contrast to prior works, we find vanilla Transformers to match the performance of S4 on Long Range Arena when properly pretrained, and we improve the best reported results of SSMs on the PathX-256 task by 20 absolute points. Subsequently, we analyze the utility of previously-proposed structured parameterizations for SSMs and show they become mostly redundant in the presence of data-driven initialization obtained through pretraining. Our work shows that, when evaluating different architectures on supervised tasks, incorporation of data-driven priors via pretraining is essential for reliable performance estimation, and can be done efficiently.

Protein Discovery with Discrete Walk-Jump Sampling
Nathan C. Frey, Dan Berenberg, Karina Zadorozhny, Joseph Kleinhenz, Julien Lafrance-Vanasse, Isidro Hotzel, Yan Wu, Stephen Ra, Richard Bonneau, Kyunghyun Cho, Andreas Loukas, Vladimir Gligorijevic, Saeed Saremi

Abstract: We resolve difficulties in training and sampling from a discrete generative model by learning a smoothed energy function, sampling from the smoothed data manifold with Langevin Markov chain Monte Carlo (MCMC), and projecting back to the true data manifold with one-step denoising. Our Discrete Walk-Jump Sampling formalism combines the contrastive divergence training of an energy-based model and improved sample quality of a score-based model, while simplifying training and sampling by requiring only a single noise level. We evaluate the robustness of our approach on generative modeling of antibody proteins and introduce the distributional conformity score to benchmark protein generative models. By optimizing and sampling from our models for the proposed distributional conformity score, 97-100% of generated samples are successfully expressed and purified and 70% of functional designs show equal or improved binding affinity compared to known functional antibodies on the first attempt in a single round of laboratory experiments. We also report the first demonstration of long-run fast-mixing MCMC chains where diverse antibody protein classes are visited in a single MCMC chain.

Vision Transformers Need Registers
Timothée Darcet, Maxime Oquab, Julien Mairal, Piotr Bojanowski

Abstract: Transformers have recently emerged as a powerful tool for learning visual representations. In this paper, we identify and characterize artifacts in feature maps of both supervised and self-supervised ViT networks. The artifacts correspond to high-norm tokens appearing during inference primarily in low-informative background areas of images, that are repurposed for internal computations. We propose a simple yet effective solution based on providing additional tokens to the input sequence of the Vision Transformer to fill that role. We show that this solution fixes that problem entirely for both supervised and self-supervised models, sets a new state of the art for self-supervised visual models on dense visual prediction tasks, enables object discovery methods with larger models, and most importantly leads to smoother feature maps and attention maps for downstream visual processing.

Honourable mentions

Amortizing intractable inference in large language models
Edward J Hu, Moksh Jain, Eric Elmoznino, Younesse Kaddar, Guillaume Lajoie, Yoshua Bengio, Nikolay Malkin

Approximating Nash Equilibria in Normal-Form Games via Stochastic Optimization
Ian Gemp, Luke Marris, Georgios Piliouras

Beyond Weisfeiler-Lehman: A Quantitative Framework for GNN Expressiveness
Bohang Zhang, Jingchu Gai, Yiheng Du, Qiwei Ye, Di He, Liwei Wang

Flow Matching on General Geometries
Ricky T. Q. Chen, Yaron Lipman

Is ImageNet worth 1 video? Learning strong image encoders from 1 long unlabelled video
Shashanka Venkataramanan, Mamshad Nayeem Rizve, Joao Carreira, Yuki M Asano, Yannis Avrithis

Meta Continual Learning Revisited: Implicitly Enhancing Online Hessian Approximation via Variance Reduction
Yichen Wu, Long-Kai Huang, Renzhen Wang, Deyu Meng, Ying Wei

Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs
Suyu Ge, Yunan Zhang, Liyuan Liu, Minjia Zhang, Jiawei Han, Jianfeng Gao

Proving Test Set Contamination in Black-Box Language Models
Yonatan Oren, Nicole Meister, Niladri S. Chatterji, Faisal Ladhak, Tatsunori Hashimoto

Robust agents learn causal world models
Jonathan Richens, Tom Everitt

The mechanistic basis of data dependence and abrupt learning in an in-context classification task
Gautam Reddy

Towards a statistical theory of data selection under weak supervision
Germain Kolossov, Andrea Montanari, Pulkit Tandon

Awards committee for the outstanding paper awards: Eunsol Choi, Katja Hofmann, Ming-Yu Liu, Nan Jiang, Stephan Günnemann, Suvrit Sra, Thomas Kipf, Volkan Cevher

You can find out more about the selection process here.

Lucy Smith , Managing Editor for AIhub.

Lucy Smith , Managing Editor for AIhub.