Lora Fanda

Lora Fanda


Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
7 March 2022 in Nature Neuroscience
Verified
Journal Article
Open access
PDF file
AGAriel Goldstein
Ariel Goldstein
ZZZaid Zada
Zaid Zada
EBEliav Buchnik
Eliav Buchnik
MSMariano Schain
Mariano Schain
APAmy Price
Amy Price
UHUri Hasson
Uri Hasson
Departing from traditional linguistic models, advances in deep learning have resulted in a new type of predictive (autoregressive) deep language models (DLMs). Using a self-supervised next-word prediction task, these models generate appropriate linguistic responses in a given context. In the current study, nine participants listened to a 30-min podcast while their brain responses were recorded using electrocorticography (ECoG). We provide empirical evidence that the human brain and autoregressive DLMs share three fundamental computational principles as they process the same natural narrative: (1) both are engaged in continuous next-word prediction before word onset; (2) both match their pre-onset predictions to the incoming word to calculate post-onset surprise; (3) both rely on contextual embeddings to represent words in natural contexts. Together, our findings suggest that autoregressive DLMs provide a new and biologically feasible computational framework for studying the neural basis of language.

24 November 2020 in Neurology
Verified
Journal Article
Open access
PDF file
JSJennifer Shum
Jennifer Shum
LFLora Fanda
Lora Fanda
PDPatricia Dugan
Patricia Dugan
WDWerner K. Doyle
Werner K. Doyle
ODOrrin Devinsky
Orrin Devinsky
AFAdeen Flinker
Adeen Flinker
Objective The combined spatiotemporal dynamics underlying sign language production remain largely unknown. To investigate these dynamics compared to speech production, we used intracranial electrocorticography during a battery of language tasks. Methods We report a unique case of direct cortical surface recordings obtained from a neurosurgical patient with intact hearing who is bilingual in English and American Sign Language. We designed a battery of cognitive tasks to capture multiple modalities of language processing and production. Results We identified 2 spatially distinct cortical networks: ventral for speech and dorsal for sign production. Sign production recruited perirolandic, parietal, and posterior temporal regions, while speech production recruited frontal, perisylvian, and perirolandic regions. Electrical cortical stimulation confirmed this spatial segregation, identifying mouth areas for speech production and limb areas for sign production. The temporal dynamics revealed superior parietal cortex activity immediately before sign production, suggesting its role in planning and producing sign language. Conclusions Our findings reveal a distinct network for sign language and detail the temporal propagation supporting sign production.

3 December 2020 in bioRxiv
Verified
Preprint
Open access
PDF file
AGAriel Goldstein
Ariel Goldstein
ZZZaid Zada
Zaid Zada
EBEliav Buchnik
Eliav Buchnik
MSMariano Schain
Mariano Schain
APAmy Price
Amy Price
UHUri Hasson
Uri Hasson
Departing from traditional linguistic models, advances in deep learning have resulted in a new type of predictive (autoregressive) deep language models (DLMs). Using a self-supervised next-word prediction task, these models are trained to generate appropriate linguistic responses in a given context. We provide empirical evidence that the human brain and autoregressive DLMs share three fundamental computational principles as they process natural language: 1) both are engaged in continuous next-word prediction before word-onset; 2) both match their pre-onset predictions to the incoming word to calculate post-onset surprise (i.e., prediction error signals); 3) both represent words as a function of the previous context. In support of these three principles, our findings indicate that: a) the neural activity before word-onset contains context-dependent predictive information about forthcoming words, even hundreds of milliseconds before the words are perceived; b) the neural activity after word-onset reflects the surprise level and prediction error; and c) autoregressive DLM contextual embeddings capture the neural representation of context-specific word meaning better than arbitrary or static semantic embeddings. Together, our findings suggest that autoregressive DLMs provide a novel and biologically feasible computational framework for studying the neural basis of language.

13 December 2022 in Electronics
Verified
Journal Article
Open access
PDF file
VCVictor Contreras
Victor Contreras
NMNiccolo Marini
Niccolo Marini
LFLora Fanda
Lora Fanda
GMGaetano Manzo
Gaetano Manzo
Yazan MuallaYazan Mualla
Yazan Mualla
Davide CalvaresiDavide Calvaresi
Davide Calvaresi
Background: Despite the advancement in eXplainable Artificial Intelligence, the explanations provided by model-agnostic predictors still call for improvements (i.e., lack of accurate descriptions of predictors’ behaviors). Contribution: We present a tool for Deep Explanations and Rule Extraction (DEXiRE) to approximate rules for Deep Learning models with any number of hidden layers. Methodology: DEXiRE proposes the binarization of neural networks to induce Boolean functions in the hidden layers, generating as many intermediate rule sets. A rule set is inducted between the first hidden layer and the input layer. Finally, the complete rule set is obtained using inverse substitution on intermediate rule sets and first-layer rules. Statistical tests and satisfiability algorithms reduce the final rule set’s size and complexity (filtering redundant, inconsistent, and non-frequent rules). DEXiRE has been tested in binary and multiclass classifications with six datasets having different structures and models. Results: The performance is consistent (in terms of accuracy, fidelity, and rule length) with respect to the state-of-the-art rule extractors (i.e., ECLAIRE). Moreover, compared with ECLAIRE, DEXiRE has generated shorter rules (i.e., up to 74% fewer terms) and has shortened the execution time (improving up to 197% in the best-case scenario). Conclusions: DEXiRE can be applied for binary and multiclass classification of deep learning predictors with any number of hidden layers. Moreover, DEXiRE can identify the activation pattern per class and use it to reduce the search space for rule extractors (pruning irrelevant/redundant neurons)—shorter rules and execution times with respect to ECLAIRE.