Related papers: Introduction to the Report "Interlanguages and Syn…
We present a feedforward graph architecture in which heterogeneous frozen large language models serve as computational nodes, communicating through a shared continuous latent space via learned linear projections. Building on recent work…
The proliferation of large-scale and structurally complex data has spurred the integration of machine learning methods into statistical modeling. Recurrent neural networks (RNNs), a foundational class of models for time-dependent data, can…
Dataflow programming is a popular and convenient programming paradigm in systems modelling, optimisation, and machine learning. It has a number of advantages, for instance the lacks of control flow allows computation to be carried out in…
Reinforcement learning (RL) has become a pivotal technology in the post-training phase of large language models (LLMs). Traditional task-colocated RL frameworks suffer from significant scalability bottlenecks, while task-separated RL…
This thesis concerns the implementation of Lambda Prolog, a higher-order logic programming language that supports the lambda-tree syntax approach to representing and manipulating formal syntactic objects. Lambda Prolog achieves its…
End-to-end models are gaining wider attention in the field of automatic speech recognition (ASR). One of their advantages is the simplicity of building that directly recognizes the speech frame sequence into the text label sequence by…
Heterogeneous many-cores are now an integral part of modern computing systems ranging from embedding systems to supercomputers. While heterogeneous many-core design offers the potential for energy-efficient high-performance, such potential…
In the case of multi-threading as found in contemporary programming languages, parallel processes are interleaved according to what is known as a process-scheduling policy in the field of operating systems. In a previous paper, we extend…
The rapid evolution of Large Language Models (LLMs) towards long-context reasoning and sparse architectures has pushed memory requirements far beyond the capacity of individual device HBM. While emerging supernode architectures offer…
Generating text from structured data is challenging because it requires bridging the gap between (i) structure and natural language (NL) and (ii) semantically underspecified input and fully specified NL output. Multilingual generation…
The proposed framework provides a general model of concurrent imperative programming. Programs are modeled as formal languages and concurrency as an interleaving (or shuffle) operator. This yields a simple and elegant algebra of programs.…
A new class of parallel algorithms is introduced that can achieve a complexity of O(n^3/2) with respect to the interprocessor communication, in the exact computation of systems with pairwise mutual interactions of all elements. Hitherto,…
Parallel computing has established itself as another standard method for applied research and data analysis. The R system, being internally constrained to mostly singly-threaded operations, can nevertheless be used along with different…
Despite real-time planners exhibiting remarkable performance in autonomous driving, the growing exploration of Large Language Models (LLMs) has opened avenues for enhancing the interpretability and controllability of motion planning.…
We introduce a framework for designing multi-scale, adaptive, shift-invariant frames and bi-frames for representing signals. The new framework, called AdaFrame, improves over dictionary learning-based techniques in terms of computational…
Existing Speech Language Model (SLM) scaling analysis paints a bleak picture. It predicts that SLMs require much more compute and data compared to text, leading some to question the feasibility of training high-quality SLMs. However, modern…
RAPID-LLM is a unified performance modeling framework for large language model (LLM) training and inference on GPU clusters. It couples a DeepFlow-based frontend that generates hardware-aware, operator-level Chakra execution traces from an…
Time is a crucial factor in modelling dynamic behaviours of intelligent agents: activities have a determined temporal duration in a real-world environment, and previous actions influence agents' behaviour. In this paper, we propose a…
Planning is a fundamental property of human intelligence. Reasoning about asynchronous plans is challenging since it requires sequential and parallel planning to optimize time costs. Can large language models (LLMs) succeed at this task?…
Reactive languages are dedicated to the programming of systems which interact continuously and concurrently with their environment. Values take the form of unbounded streams modeling the (discrete) passing of time or the sequence of…