Daniel N. Rockmore

Neel Guha, Julian Nyarko, Daniel E. Ho et al.

The advent of large language models (LLMs) and their adoption by the legal community has given rise to the question: what types of legal reasoning can LLMs perform? To enable greater study of this question, we present LegalBench: a collaboratively constructed legal reasoning benchmark consisting of 162 tasks covering six different types of legal reasoning. LegalBench was built through an interdisciplinary process, in which we collected tasks designed and hand-crafted by legal professionals. Because these subject matter experts took a leading role in construction, tasks either measure legal reasoning capabilities that are practically useful, or measure reasoning skills that lawyers find interesting. To enable cross-disciplinary conversations about LLMs in the law, we additionally show how popular legal frameworks for describing legal reasoning -- which distinguish between its many forms -- correspond to LegalBench tasks, thus giving lawyers and LLM developers a common vocabulary. This paper describes LegalBench, presents an empirical evaluation of 20 open-source and commercial LLMs, and illustrates the types of research explorations LegalBench enables.

David Maslen, Daniel N. Rockmore, Sarah Wolff

We present a general diagrammatic approach to the construction of efficient algorithms for computing the Fourier transform of a function on a finite group. By extending work which connects Bratteli diagrams to the construction of Fast Fourier Transform algorithms %\cite{sovi}, we make explicit use of the path algebra connection to the construction of Gel'fand-Tsetlin bases and work in the setting of quivers. We relate this framework to the construction of a {\em configuration space} derived from a Bratteli diagram. In this setting the complexity of an algorithm for computing a Fourier transform reduces to the calculation of the dimension of the associated configuration space. Our methods give improved upper bounds for computing the Fourier transform for the general linear groups over finite fields, the classical Weyl groups, and homogeneous spaces of finite groups, while also recovering the best known algorithms for the symmetric group and compact Lie groups.

Martin Malandro, Daniel N. Rockmore

We define the notion of the Fourier transform for the rook monoid (also called the symmetric inverse semigroup) and provide two efficient divide-and-conquer algorithms (fast Fourier transforms, or FFTs) for computing it. This paper marks the first extension of group FFTs to non-group semigroups.

Wooram Park, Daniel N. Rockmore, Dean Madden et al.

This paper addresses the deconvolution of an image that has been obtained by superimposing many copies of an underlying unknown image of interest. The superposition is assumed to not be exact due to noise, and is described using an error distribution in position, orientation, or both. We assume that a good estimate of the error distribution is known. The most natural approach to take for the purely translational case is to apply the Fourier transform and use the classical convolution theorem together with a Weiner filter to invert. In the case of purely rotational deblurring, the similar Fourier analysis is applied. That is, for an blurred image function defined on polar coordinates, the Fourier series and the convolution theorem for the series can be applied. In the case of combinations of translational and rotational errors, the motion-group Fourier transform is used. In addition, for the three cases we present the alternative method using Hermite and Laguerre-Fourier expansion, which has a special property in Fourier transform. The problem that is solved here is motivated by one of the steps in the cryo-electron-tomographic reconstruction of biomolecular complexes such as viruses and ion channels.

Harrison F. Stropkay, Jiayi Chen, Mohammad J. Latifi et al.

We show that large language models (LLMs) can be used to distinguish the writings of different authors. Specifically, an individual GPT-2 model, trained from scratch on the works of one author, will predict held-out text from that author more accurately than held-out text from other authors. We suggest that, in this way, a model trained on one author's works embodies the unique writing style of that author. We first demonstrate our approach on books written by eight different (known) authors. We also use this approach to confirm R. P. Thompson's authorship of the well-studied 15th book of the Oz series, originally attributed to F. L. Baum.

Daniel N. Rockmore, Chen Fang, Nicholas J. Foti et al.

We explore how ideas from infectious disease and genetics can be used to uncover patterns of cultural inheritance and innovation in a corpus of 591 national constitutions spanning 1789 - 2008. Legal "Ideas" are encoded as "topics" - words statistically linked in documents - derived from topic modeling the corpus of constitutions. Using these topics we derive a diffusion network for borrowing from ancestral constitutions back to the US Constitution of 1789 and reveal that constitutions are complex cultural recombinants. We find systematic variation in patterns of borrowing from ancestral texts and "biological"-like behavior in patterns of inheritance with the distribution of "offspring" arising through a bounded preferential-attachment process. This process leads to a small number of highly innovative (influential) constitutions some of which have yet to have been identified as so in the current literature. Our findings thus shed new light on the critical nodes of the constitution-making network. The constitutional network structure reflects periods of intense constitution creation, and systematic patterns of variation in constitutional life-span and temporal influence.

Chen Fang, Daniel N. Rockmore

Multi-task learning (MTL) improves prediction performance in different contexts by learning models jointly on multiple different, but related tasks. Network data, which are a priori data with a rich relational structure, provide an important context for applying MTL. In particular, the explicit relational structure implies that network data is not i.i.d. data. Network data also often comes with significant metadata (i.e., attributes) associated with each entity (node). Moreover, due to the diversity and variation in network data (e.g., multi-relational links or multi-category entities), various tasks can be performed and often a rich correlation exists between them. Learning algorithms should exploit all of these additional sources of information for better performance. In this work we take a metric-learning point of view for the MTL problem in the network context. Our approach builds on structure preserving metric learning (SPML). In particular SPML learns a Mahalanobis distance metric for node attributes using network structure as supervision, so that the learned distance function encodes the structure and can be used to predict link patterns from attributes. SPML is described for single-task learning on single network. Herein, we propose a multi-task version of SPML, abbreviated as MT-SPML, which is able to learn across multiple related tasks on multiple networks via shared intermediate parametrization. MT-SPML learns a specific metric for each task and a common metric for all tasks. The task correlation is carried through the common metric and the individual metrics encode task specific information. When combined together, they are structure-preserving with respect to individual tasks. MT-SPML works on general networks, thus is suitable for a wide variety of problems. In experiments, we challenge MT-SPML on two real-word problems, where MT-SPML achieves significant improvement.