Emilie Josephs

Emilie Josephs

Auto-generated profile, based on publicly available information

Emilie Josephs

Department of Psychology, Harvard University

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

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

Representative Publications

View All Publications
28 February 2024 in Journal of Online Trust and Safety
Verified
Journal Article
Open access
PDF file
EJEmilie Josephs
Emilie Josephs
CFCamilo Fosco
Camilo Fosco
AOAude Oliva
Aude Oliva
The increasing reach of deepfakes raises practical questions about people’s ability to detect false videos online. How vulnerable are people to deepfake videos? What technologies can help improve detection? Previous experiments that measure human deepfake detection historically omit a number of conditions that can exist in typical browsing conditions. Here, we operationalized four such conditions (low prevalence, brief presentation, low video quality, and divided attention), and found in a series of online experiments that all conditions lowered detection relative to baseline, suggesting that the current literature underestimates people’s susceptibility to deepfakes. Next, we examined how AI assistance could be integrated into the human decision process. We found that a model that exposes deepfakes by amplifying artifacts increases detection rates, and also leads to higher rates of incorporating AI feedback and higher final confidence than text-based prompts. Overall, this suggests that visual indicators that cause distortions on fake videos may be effective at mitigating the impact of falsified video.

1 September 2015 in Journal of Vision
Verified
Journal Article
Open access
EJEmilie Josephs
Emilie Josephs
MCMatthew Cain
Matthew Cain
BHBarbara Hidalgo-Sotelo
Barbara Hidalgo-Sotelo
GCGregory Cook
Gregory Cook
NCNelson Chang
Nelson Chang
JWJeremy Wolfe
Jeremy Wolfe
Does stereoscopic information improve visual search? We know that attention can be guided efficiently by stereopsis, for example, to the near target among far distractors (Nakayama and Silverman, 1986), but, when searching through a real scene, does it help if that scene is presented stereoscopically? Certainly, scenes appear to be more vividly real in 3D. However, we present three experiments in which the addition of stereo did not alter scene search very much. In Experiment 1, 12 observers searched twice for 10 target objects in each of 18 photographic scenes (a ‘repeated’ search task). Note that observers were not cued to search for a target at a specific depth, stereopsis simply added to the vividness of the scene. Reaction time and accuracy did not differ significantly in stereoscopic and monoscopic conditions. In Experiment 2, using similar images, we found no differences between 2D and 3D conditions on time to first fixation on the target or on the average length of saccades. However, gaze durations were significantly shorter in 3D scenes. Since gaze durations are typically taken to measure processing time, this may suggest that it was easier to disambiguate surfaces and/or objects in 3D, although this advantage did not translate to benefits in search times. In a final experiment, we reduced the stimulus array to a set of colored rendered objects distributed in depth against a plain background. In this task, the addition of stereo produced shorter reaction times, even though stereo information was not predictive of target location. Our real scenes may have contained such a rich array of cues to target location that the addition of stereo may not have contributed much additional information. It may be in more difficult searches, including more challenging real world tasks, that stereopsis will be an asset. Meeting abstract presented at VSS 2015

6 January 2022 in bioRxiv
Verified
Preprint
Open access
PDF file
JPJeongho Park
Jeongho Park
EJEmilie Josephs
Emilie Josephs
TKTalia Konkle
Talia Konkle
We can easily perceive the spatial scale depicted in a picture, regardless of whether it is a small space (e.g., a close-up view of a chair) or a much larger space (e.g., an entire class room). How does the human visual system encode this continuous dimension? Here, we investigated the underlying neural coding of depicted spatial scale, by examining the voxel tuning and topographic organization of brain responses. We created naturalistic yet carefully-controlled stimuli by constructing virtual indoor environments, and rendered a series of snapshots to smoothly sample between a close-up view of the central object and far-scale view of the full environment (object-to-scene continuum). Human brain responses were measured to each position using functional magnetic resonance imaging. We did not find evidence for a smooth topographic mapping for the object-to-scene continuum on the cortex. Instead, we observed large swaths of cortex with opposing ramp-shaped profiles, with highest responses to one end of the object-to-scene continuum or the other, and a small region showing a weak tuning to intermediate scale views. Importantly, when we considered the multi-voxel patterns of the entire ventral occipito-temporal cortex, we found smooth and linear representation of the object-to-scene continuum. Thus, our results together suggest that depicted spatial scale is coded parametrically in large-scale population codes across the entire ventral occipito-temporal cortex.