Modern neural recording techniques such as two-photon imaging or Neuropixel probes allow to acquire vast time-series datasets with responses of hundreds or thousands of neurons. Contrastive learning is a powerful self-supervised framework for learning representations of complex datasets. Existing applications for neural time series rely on generic data augmentations and do not exploit the multi-trial data structure inherent in many neural datasets. Here we present TRACE, a new contrastive learning framework that averages across different subsets of trials to generate positive pairs. TRACE allows to directly learn a two-dimensional embedding, combining ideas from contrastive learning and neighbor embeddings. We show that TRACE outperforms other methods, resolving fine response differences in simulated data. Further, using in vivo recordings, we show that the representations learned by TRACE capture both biologically relevant continuous variation, cell-type-related cluster structure, and can assist data quality control.

Pairwise distance-based costs are crucial for self-supervised and contrastive feature learning. Mixture Density Networks (MDNs) are a widely used approach for generative models and density approximation, using neural networks to produce multiple centers that define a Gaussian mixture. By combining MDNs with contrastive costs, this paper proposes data density approximation using four types of kernelized matrix costs: the scalar cost, the vector-matrix cost, the matrix-matrix cost (the trace of Schur complement), and the SVD cost (the nuclear norm), for learning multiple centers required to define a mixture density.

We study a class of optimization problems motivated by automating the design and update of AI systems like coding assistants, robots, and copilots. AutoDiff frameworks, like PyTorch, enable efficient end-to-end optimization of differentiable systems. However, general computational workflows can be non-differentiable and involve rich feedback (e.g. We investigate end-to-end generative optimization – using generative models such as LLMs within the optimizer for automatic updating of general computational workflows. We discover that workflow execution traces are akin to back-propagated gradients in AutoDiff and can provide key information to interpret feedback for efficient optimization.

Causal inference can be formalized as Bayesian inference that combines a prior distribution over causal models and likelihoods that account for both observations and interventions. We show that it is possible to implement this approach using a sufficiently expressive probabilistic programming language. Priors are represented using probabilistic programs that generate source code in a domain specific language. Interventions are represented using probabilistic programs that edit this source code to modify the original generative process. This approach makes it straightforward to incorporate data from atomic interventions, as well as shift interventions, variance-scaling interventions, and other interventions that modify causal structure. This approach also enables the use of general-purpose inference machinery for probabilistic programs to infer probable causal structures and parameters from data. This abstract describes a prototype of this approach in the Gen probabilistic programming language.

In general, modeling is a complex and creative task, and building qualitative models is no exception. One way of automating this task is by means of machine learning. Observed behaviors of a modeled system are used as examples for a learning algorithm that constructs a model that is consistent with the data. In this article, we review approaches to learning qualitative models, either from numeric data or qualitative observations. However, an important practical question is how do we construct qualitative models in the first place.

A key element of transfer learning is to identify structured knowledge to enable the knowledge transfer. Structured knowledge comes in different forms, depending on the nature of the learning problem and characteristics of the domains. In this article, we describe three of our recent works on transfer learning in a progressively more sophisticated order of the structured knowledge being transferred. We show that optimization methods and techniques inspired by the concerns of data reuse can be applied to extract and transfer deep structural knowledge between a variety of source and target problems. This often happens when we meet with new domains and encounter new tasks.

Recurrent neural networks have been very successful at predicting sequences of words in tasks such as language modeling. However, all such models are based on the conventional classification framework, where the model is trained against one-hot targets, and each word is represented both as an input and as an output in isolation. This causes inefficiencies in learning both in terms of utilizing all of the information and in terms of the number of parameters needed to train. We introduce a novel theoretical framework that facilitates better learning in language modeling, and show that our framework leads to tying together the input embedding and the output projection matrices, greatly reducing the number of trainable variables. Our framework leads to state of the art performance on the Penn Treebank with a variety of network models.

Approximating a matrix by a small subset of its columns is a known problem in numerical linear algebra. Algorithms that address this problem have been used in areas which include, among others, sparse approximation, unsupervised feature selection, data mining, and knowledge representation. Such algorithms were investigated since the 1960's, with recent results that use randomization. The problem is believed to be NP-Hard, and to the best of our knowledge there are no previously published algorithms aimed at computing optimal solutions. We show how to model the problem as a graph search, and propose a heuristic based on eigenvalues of related matrices. Applying the A* search strategy with this heuristic is guaranteed to find the optimal solution. Experimental results on common datasets show that the proposed algorithm can effectively select columns from moderate size matrices, typically improving by orders of magnitude the run time of exhaustive search. We also show how to combine the proposed algorithm with other non-optimal (but much faster) algorithms in a ``two stage'' framework, which is guaranteed to improve the accuracy of the other algorithms.

It provides inference and explanation facilities, and incorporates a novel form of plausible inference. YAPES is a specialized interpreter for logic programs. Figure 1 illustrates its top level structure. A PROLOG interpreter (or compiler) executes such programs consisting of sets of Horn clauses, a form of first-order logic. The YAPES system also executes such programs, as well as programs in an extended version of Horn clause logic which uses certainties as truth values, rather than just true and false.

Recent psychological experiments intend to show that social intentions can be read from the recording of motor actions (Becchio, Sartori, and Castiello 2010; Ferri et al. 2011). At the center of the debate is the hypothesis that the motor system is (Blackemore and Decety 2001), or is not (Jacob and Jeannerod 2005) used to recognize social intentions, with a potential openning to a bottom-up understanding of social behavior, agentivity and theory of mind. In (Becchio et al. 2007), the authors proposed to record the arm's trajectories during episodes of a "pick and place" task with a motor vs social outcome. The results provided evidence for differences in motor patterning depending on the social context and intention, but where not yet a direct evidence of the involvement of the motor system in recognizing social intention. In (Becchio, Sartori, and Castiello 2010; Ferri et al. 2011), the authors show how social affordances can change the movement parametrization with the hypothesis that a same action linked to a social context may involve an increase of the index of difficulty.