Zobrazeno 1 - 10
of 23
pro vyhledávání: '"Hennigen, Lucas Torroba"'
Autor:
Du, Li, Amini, Afra, Hennigen, Lucas Torroba, Yu, Xinyan Velocity, Eisner, Jason, Lee, Holden, Cotterell, Ryan
Recent papers have demonstrated the possibility of energy-based text generation by adapting gradient-based sampling algorithms, a paradigm of MCMC algorithms that promises fast convergence. However, as we show in this paper, previous attempts on this
Externí odkaz:
http://arxiv.org/abs/2312.17710
Autor:
Hennigen, Lucas Torroba, Shen, Shannon, Nrusimha, Aniruddha, Gapp, Bernhard, Sontag, David, Kim, Yoon
LLMs are vulnerable to hallucinations, and thus their outputs generally require laborious human verification for high-stakes applications. To this end, we propose symbolically grounded generation (SymGen) as a simple approach for enabling easier manu
Externí odkaz:
http://arxiv.org/abs/2311.09188
Many popular feature-attribution methods for interpreting deep neural networks rely on computing the gradients of a model's output with respect to its inputs. While these methods can indicate which input features may be important for the model's pred
Externí odkaz:
http://arxiv.org/abs/2307.03056
Autor:
Hennigen, Lucas Torroba, Kim, Yoon
Masked language models (MLM) do not explicitly define a distribution over language, i.e., they are not language models per se. However, recent work has implicitly treated them as such for the purposes of generation and scoring. This paper studies met
Externí odkaz:
http://arxiv.org/abs/2305.15501
Autor:
Wang, Peihao, Panda, Rameswar, Hennigen, Lucas Torroba, Greengard, Philip, Karlinsky, Leonid, Feris, Rogerio, Cox, David Daniel, Wang, Zhangyang, Kim, Yoon
Scaling transformers has led to significant breakthroughs in many domains, leading to a paradigm in which larger versions of existing models are trained and released on a periodic basis. New instances of such models are typically trained completely f
Externí odkaz:
http://arxiv.org/abs/2303.00980
Autor:
Du, Li, Hennigen, Lucas Torroba, Pimentel, Tiago, Meister, Clara, Eisner, Jason, Cotterell, Ryan
Language modeling, a central task in natural language processing, involves estimating a probability distribution over strings. In most cases, the estimated distribution sums to 1 over all finite strings. However, in some pathological cases, probabili
Externí odkaz:
http://arxiv.org/abs/2212.10502
Autor:
Stoehr, Niklas, Hennigen, Lucas Torroba, Valvoda, Josef, West, Robert, Cotterell, Ryan, Schein, Aaron
Measuring the intensity of events is crucial for monitoring and tracking armed conflict. Advances in automated event extraction have yielded massive data sets of "who did what to whom" micro-records that enable data-driven approaches to monitoring co
Externí odkaz:
http://arxiv.org/abs/2210.03971
Autor:
Batsuren, Khuyagbaatar, Goldman, Omer, Khalifa, Salam, Habash, Nizar, Kieraś, Witold, Bella, Gábor, Leonard, Brian, Nicolai, Garrett, Gorman, Kyle, Ate, Yustinus Ghanggo, Ryskina, Maria, Mielke, Sabrina J., Budianskaya, Elena, El-Khaissi, Charbel, Pimentel, Tiago, Gasser, Michael, Lane, William, Raj, Mohit, Coler, Matt, Samame, Jaime Rafael Montoya, Camaiteri, Delio Siticonatzi, Sagot, Benoît, Rojas, Esaú Zumaeta, Francis, Didier López, Oncevay, Arturo, Bautista, Juan López, Villegas, Gema Celeste Silva, Hennigen, Lucas Torroba, Ek, Adam, Guriel, David, Dirix, Peter, Bernardy, Jean-Philippe, Scherbakov, Andrey, Bayyr-ool, Aziyana, Anastasopoulos, Antonios, Zariquiey, Roberto, Sheifer, Karina, Ganieva, Sofya, Cruz, Hilaria, Karahóǧa, Ritván, Markantonatou, Stella, Pavlidis, George, Plugaryov, Matvey, Klyachko, Elena, Salehi, Ali, Angulo, Candy, Baxi, Jatayu, Krizhanovsky, Andrew, Krizhanovskaya, Natalia, Salesky, Elizabeth, Vania, Clara, Ivanova, Sardana, White, Jennifer, Maudslay, Rowan Hall, Valvoda, Josef, Zmigrod, Ran, Czarnowska, Paula, Nikkarinen, Irene, Salchak, Aelita, Bhatt, Brijesh, Straughn, Christopher, Liu, Zoey, Washington, Jonathan North, Pinter, Yuval, Ataman, Duygu, Wolinski, Marcin, Suhardijanto, Totok, Yablonskaya, Anna, Stoehr, Niklas, Dolatian, Hossep, Nuriah, Zahroh, Ratan, Shyam, Tyers, Francis M., Ponti, Edoardo M., Aiton, Grant, Arora, Aryaman, Hatcher, Richard J., Kumar, Ritesh, Young, Jeremiah, Rodionova, Daria, Yemelina, Anastasia, Andrushko, Taras, Marchenko, Igor, Mashkovtseva, Polina, Serova, Alexandra, Prud'hommeaux, Emily, Nepomniashchaya, Maria, Giunchiglia, Fausto, Chodroff, Eleanor, Hulden, Mans, Silfverberg, Miikka, McCarthy, Arya D., Yarowsky, David, Cotterell, Ryan, Tsarfaty, Reut, Vylomova, Ekaterina
The Universal Morphology (UniMorph) project is a collaborative effort providing broad-coverage instantiated normalized morphological inflection tables for hundreds of diverse world languages. The project comprises two major thrusts: a language-indepe
Externí odkaz:
http://arxiv.org/abs/2205.03608
Autor:
Stańczak, Karolina, Ponti, Edoardo, Hennigen, Lucas Torroba, Cotterell, Ryan, Augenstein, Isabelle
The success of multilingual pre-trained models is underpinned by their ability to learn representations shared by multiple languages even in absence of any explicit supervision. However, it remains unclear how these models learn to generalise across
Externí odkaz:
http://arxiv.org/abs/2205.02023
Autor:
Stańczak, Karolina, Hennigen, Lucas Torroba, Williams, Adina, Cotterell, Ryan, Augenstein, Isabelle
The success of pre-trained contextualized representations has prompted researchers to analyze them for the presence of linguistic information. Indeed, it is natural to assume that these pre-trained representations do encode some level of linguistic k
Externí odkaz:
http://arxiv.org/abs/2201.08214