Search

Job Corps is the largest and most comprehensive education and job training program in the United States for young people ages 16 to 24 who are not in school or working. Most students live at and take classes at its approximately 120 residential centers. To explore ways to improve student outcomes and reduce per-student costs, Job Corps pilot tested two nonresidential models where students received education and training at colleges, supported by intensive personal and career counseling, instead of at a center. The two pilot programs were: (1) The Job Corps Scholars (JCS) pilot program consisted of 26 colleges each delivering its own version of the Job Corps program to eligible young people; and (2) In the Idaho Job Corps (IJC) pilot program, Idaho paired its Boise Job Corps center with a nearby community college and opened three addition fully nonresidential programs at other colleges. The two pilot programs together offered 30 diverse examples of college-partnered Job Corps programs to study. The implementation and outcomes study focused on: (1) The types of students who enrolled in these programs; (2) The ability of colleges to serve young people eligible for Job Corps, who typically face more barriers than other students; and (3) The outcomes achieved by students in the pilot programs.

Across the country, as technology continues to advance rapidly, the labor market exhibits a growing need for workers who receive frequent and ongoing skill development. Employers in many fields struggle to find adequately trained workers to meet their needs. Community college noncredit career and technical education (CTE) programs are an important contributor to skill and workforce development and help to close this "skills gap." This brief summarizes early findings from a study of FastForward, which uses a pay-for-performance model to fund noncredit CTE programs at the 23 colleges in the Virginia Community College System. FastForward aims to increase the supply of workers who receive credentials for high-demand occupations in Virginia. The FastForward study is focused on identifying institutional and programmatic factors that may influence learners' academic and labor market outcomes. This brief presents findings on the different approaches used by colleges and programs to deliver training, student and staff experiences in these CTE programs, and students' academic and labor market outcomes.

The orthogonal momentum amplituhedron O_k was introduced simultaneously in 2021 by Huang, Kojima, Wen, and Zhang, and by He, Kuo, and Zhang, in the study of scattering amplitudes of ABJM theory. It was conjectured that it admits a decomposition into BCFW cells. We prove this conjecture.

Two substantial technological advances have reshaped the public square in recent decades: first with the advent of the internet and second with the recent introduction of large language models (LLMs). LLMs offer opportunities for a paradigm shift towards more decentralized, participatory online spaces that can be used to facilitate deliberative dialogues at scale, but also create risks of exacerbating societal schisms. Here, we explore four applications of LLMs to improve digital public squares: collective dialogue systems, bridging systems, community moderation, and proof-of-humanity systems. Building on the input from over 70 civil society experts and technologists, we argue that LLMs both afford promising opportunities to shift the paradigm for conversations at scale and pose distinct risks for digital public squares. We lay out an agenda for future research and investments in AI that will strengthen digital public squares and safeguard against potential misuses of AI., Comment: 40 pages, 5 figures

Inverse folding models play an important role in structure-based design by predicting amino acid sequences that fold into desired reference structures. Models like ProteinMPNN, a message-passing encoder-decoder model, are trained to reliably produce new sequences from a reference structure. However, when applied to peptides, these models are prone to generating repetitive sequences that do not fold into the reference structure. To address this, we fine-tune ProteinMPNN to produce diverse and structurally consistent peptide sequences via Direct Preference Optimization (DPO). We derive two enhancements to DPO: online diversity regularization and domain-specific priors. Additionally, we develop a new understanding on improving diversity in decoder models. When conditioned on OpenFold generated structures, our fine-tuned models achieve state-of-the-art structural similarity scores, improving base ProteinMPNN by at least 8%. Compared to standard DPO, our regularized method achieves up to 20% higher sequence diversity with no loss in structural similarity score., Comment: Preprint. 10 pages plus appendices

Crafting a single, versatile physics-based controller that can breathe life into interactive characters across a wide spectrum of scenarios represents an exciting frontier in character animation. An ideal controller should support diverse control modalities, such as sparse target keyframes, text instructions, and scene information. While previous works have proposed physically simulated, scene-aware control models, these systems have predominantly focused on developing controllers that each specializes in a narrow set of tasks and control modalities. This work presents MaskedMimic, a novel approach that formulates physics-based character control as a general motion inpainting problem. Our key insight is to train a single unified model to synthesize motions from partial (masked) motion descriptions, such as masked keyframes, objects, text descriptions, or any combination thereof. This is achieved by leveraging motion tracking data and designing a scalable training method that can effectively utilize diverse motion descriptions to produce coherent animations. Through this process, our approach learns a physics-based controller that provides an intuitive control interface without requiring tedious reward engineering for all behaviors of interest. The resulting controller supports a wide range of control modalities and enables seamless transitions between disparate tasks. By unifying character control through motion inpainting, MaskedMimic creates versatile virtual characters. These characters can dynamically adapt to complex scenes and compose diverse motions on demand, enabling more interactive and immersive experiences., Comment: ACM Transactions on Graphics (Proc. SIGGRAPH Asia 2024) Project page: https://research.nvidia.com/labs/par/maskedmimic/

Advanced AI systems capable of generating humanlike text and multimodal content are now widely available. In this paper, we discuss the impacts that generative artificial intelligence may have on democratic processes. We consider the consequences of AI for citizens' ability to make informed choices about political representatives and issues (epistemic impacts). We ask how AI might be used to destabilise or support democratic mechanisms like elections (material impacts). Finally, we discuss whether AI will strengthen or weaken democratic principles (foundational impacts). It is widely acknowledged that new AI systems could pose significant challenges for democracy. However, it has also been argued that generative AI offers new opportunities to educate and learn from citizens, strengthen public discourse, help people find common ground, and to reimagine how democracies might work better., Comment: 25 pages

In multiplayer, first-person shooter games like Counter-Strike: Global Offensive (CS:GO), coordinated movement is a critical component of high-level strategic play. However, the complexity of team coordination and the variety of conditions present in popular game maps make it impractical to author hand-crafted movement policies for every scenario. We show that it is possible to take a data-driven approach to creating human-like movement controllers for CS:GO. We curate a team movement dataset comprising 123 hours of professional game play traces, and use this dataset to train a transformer-based movement model that generates human-like team movement for all players in a "Retakes" round of the game. Importantly, the movement prediction model is efficient. Performing inference for all players takes less than 0.5 ms per game step (amortized cost) on a single CPU core, making it plausible for use in commercial games today. Human evaluators assess that our model behaves more like humans than both commercially-available bots and procedural movement controllers scripted by experts (16% to 59% higher by TrueSkill rating of "human-like"). Using experiments involving in-game bot vs. bot self-play, we demonstrate that our model performs simple forms of teamwork, makes fewer common movement mistakes, and yields movement distributions, player lifetimes, and kill locations similar to those observed in professional CS:GO match play., Comment: The project website is at https://davidbdurst.com/mlmove/

For the successful implementation of organic electrochemical transistors in neuromorphic computing, bioelectronics, and real-time sensing applications it is essential to understand the factors that influence device switching times. Here we describe a physical-electrochemical model of the transient response to a step of the gate voltage. The model incorporates (1) ion diffusion inside the channel that governs the electronic conductivity, (2) horizontal electron transport, and (3) the external elements (capacitance, ionic resistance) of the ion dynamics in the electrolyte. We find a general expression of two different time constants that determine the vertical insertion process in terms of the kinetic parameters, in addition to the electronic transit time. We highlight the central role of the chemical capacitance in determining the modulation of the lateral conductivity. The different types of response of the drain current are classified, and we discuss the significance for synaptic operation in neuromorphic circuits. The model is confirmed by detailed simulations that enable to visualize the different ions distributions and dynamics.

Neural networks (NN) achieve remarkable results in various tasks, but lack key characteristics: interpretability, support for categorical features, and lightweight implementations suitable for edge devices. While ongoing efforts aim to address these challenges, Gradient Boosting Trees (GBT) inherently meet these requirements. As a result, GBTs have become the go-to method for supervised learning tasks in many real-world applications and competitions. However, their application in online learning scenarios, notably in reinforcement learning (RL), has been limited. In this work, we bridge this gap by introducing Gradient-Boosting RL (GBRL), a framework that extends the advantages of GBT to the RL domain. Using the GBRL framework, we implement various actor-critic algorithms and compare their performance with their NN counterparts. Inspired by shared backbones in NN we introduce a tree-sharing approach for policy and value functions with distinct learning rates, enhancing learning efficiency over millions of interactions. GBRL achieves competitive performance across a diverse array of tasks, excelling in domains with structured or categorical features. Additionally, we present a high-performance, GPU-accelerated implementation that integrates seamlessly with widely-used RL libraries (available at https://github.com/NVlabs/gbrl). GBRL expands the toolkit for RL practitioners, demonstrating the viability and promise of GBT within the RL paradigm, particularly in domains characterized by structured or categorical features.

We present SMPLOlympics, a collection of physically simulated environments that allow humanoids to compete in a variety of Olympic sports. Sports simulation offers a rich and standardized testing ground for evaluating and improving the capabilities of learning algorithms due to the diversity and physically demanding nature of athletic activities. As humans have been competing in these sports for many years, there is also a plethora of existing knowledge on the preferred strategy to achieve better performance. To leverage these existing human demonstrations from videos and motion capture, we design our humanoid to be compatible with the widely-used SMPL and SMPL-X human models from the vision and graphics community. We provide a suite of individual sports environments, including golf, javelin throw, high jump, long jump, and hurdling, as well as competitive sports, including both 1v1 and 2v2 games such as table tennis, tennis, fencing, boxing, soccer, and basketball. Our analysis shows that combining strong motion priors with simple rewards can result in human-like behavior in various sports. By providing a unified sports benchmark and baseline implementation of state and reward designs, we hope that SMPLOlympics can help the control and animation communities achieve human-like and performant behaviors., Comment: Project page: https://smplolympics.github.io/SMPLOlympics

Controlling humanoids in complex physically simulated worlds is a long-standing challenge with numerous applications in gaming, simulation, and visual content creation. In our setup, given a rich and complex 3D scene, the user provides a list of instructions composed of target locations and locomotion types. To solve this task we present PlaMo, a scene-aware path planner and a robust physics-based controller. The path planner produces a sequence of motion paths, considering the various limitations the scene imposes on the motion, such as location, height, and speed. Complementing the planner, our control policy generates rich and realistic physical motion adhering to the plan. We demonstrate how the combination of both modules enables traversing complex landscapes in diverse forms while responding to real-time changes in the environment. Video: https://youtu.be/wWlqSQlRZ9M .

Stellar abundances of cerium are of high current interest based both on observations and theoretical models, especially with regard to the neutron--magic $^{140}$Ce isotope. A large discrepancy of $s-$process stellar models relative to cerium abundance observed in globular clusters was highlighted, pointing to possible uncertainties in experimental nuclear reaction rates. In this work, the stellar neutron capture cross section of the stable cerium isotopes $^{136}$Ce, $^{138}$Ce, $^{140}$Ce, and $^{142}$Ce, were re-measured. A $^{nat}$Ce sample was irradiated with quasi-Maxwellian neutrons at $kT = 34.2$ keV using the $^{7}$Li($p,n$) reaction. The neutron field with an intensity of $3-5 \times 10^{10}$ n/s was produced by irradiating the liquid-lithium target (LiLiT) with a mA proton beam at an energy (1.92 MeV) just above the threshold at Soreq Applied Research Accelerator Facility (SARAF). The activities of the $^{nat}$Ce neutron capture products were measured using a shielded High Purity Germanium detector. Cross sections were extracted relative to that of the $^{197}$Au(n,$\gamma$) reaction and the Maxwellian-averaged cross section (MACS) of the Ce isotopes were derived. The MACS values extracted from this experiment are generally consistent with previous measurements and show for $^{140}$Ce a value $\approx 15$\% smaller than most recent experimental values., Comment: 9 pages, 12 figures. This work is supported in part by the Pazy Foundation (Israel) and the German Israeli Foundation (GIF) under Grant No. I-1500-303.7/2019. M.P. and M.F. acknowledge support by the European Union (ChETEC-INFRA)

Hysteresis in organic electrochemical transistors (OECT) is a basic effect in which the measured current depends on the voltage sweep direction and velocity. This phenomenon has an important impact on different aspects of the application of OECT, such as the switching time and the synaptic properties for neuromorphic applications. Here we address the combined ionic and electronic kinetic effects that cause the dominant hysteresis effects. We use a combination of tools consisting on basic analytical models, advanced 2D drift-diffusion simulation, and the experimental measurement of a Poly(3-hexylthiophene) (P3HT) OECT, working in an accumulation mode. We develop a general transmission line model considering drift electronic transport and ionic injection and diffusion from the electrolyte. We provide a basic classification of the transient response to a voltage pulse, according to the dominant ionic or electronic relaxation time, and the correspondent hysteresis effects of the transfer curves according to the general categories of inductive and capacitive hysteresis. These are basically related to the main control phenomenon, either the vertical diffusion of ions during doping and dedoping, or the equilibration of electronic current along the channel length.

Recent generative AI systems have demonstrated more advanced persuasive capabilities and are increasingly permeating areas of life where they can influence decision-making. Generative AI presents a new risk profile of persuasion due the opportunity for reciprocal exchange and prolonged interactions. This has led to growing concerns about harms from AI persuasion and how they can be mitigated, highlighting the need for a systematic study of AI persuasion. The current definitions of AI persuasion are unclear and related harms are insufficiently studied. Existing harm mitigation approaches prioritise harms from the outcome of persuasion over harms from the process of persuasion. In this paper, we lay the groundwork for the systematic study of AI persuasion. We first put forward definitions of persuasive generative AI. We distinguish between rationally persuasive generative AI, which relies on providing relevant facts, sound reasoning, or other forms of trustworthy evidence, and manipulative generative AI, which relies on taking advantage of cognitive biases and heuristics or misrepresenting information. We also put forward a map of harms from AI persuasion, including definitions and examples of economic, physical, environmental, psychological, sociocultural, political, privacy, and autonomy harm. We then introduce a map of mechanisms that contribute to harmful persuasion. Lastly, we provide an overview of approaches that can be used to mitigate against process harms of persuasion, including prompt engineering for manipulation classification and red teaming. Future work will operationalise these mitigations and study the interaction between different types of mechanisms of persuasion.

The amplituhedron $A_{n,k,m}$ is a geometric object introduced in the context of scattering amplitudes in $N=4$ super Yang Mills. It generalizes the positive Grassmannian (when $n=k+m$), cyclic polytopes (when $k=1$), and the bounded complex of the cyclic hyperplane arrangement (when $m=1$). Of substantial interest are the tilings of the amplituhedron, which are analogous to triangulations of a polytope. Karp, Williams and Zhang (2020) observed that the known tilings of $A_{n,k,2}$ have cardinality ${n-2 \choose k}$ and the known tilings of $A_{n,k,4}$ have cardinality the Narayana number $\frac{1}{n-3}{n-3 \choose k+1}{n-3 \choose k}$; generalizing these observations, they conjectured that for even $m$ the tilings of $A_{n, k,m}$ have cardinality the MacMahon number, the number of plane partitions which fit inside a $k \times (n-k-m) \times \frac{m}{2}$ box. We refer to this prediction as the `Magic Number Conjecture'. In this paper we prove the Magic Number Conjecture for the $m=2$ amplituhedron: that is, we show that each tiling of $A_{n,k,2}$ has cardinality ${n-2 \choose k}$. We prove this by showing that all positroid tilings of the hypersimplex $\Delta_{k+1,n}$ have cardinality ${n-2 \choose k}$, then applying T-duality. In addition, we give combinatorial necessary conditions for tiles to form a tiling of $A_{n,k,2}$; we give volume formulas for Parke-Taylor polytopes and certain positroid polytopes in terms of circular extensions of cyclic partial orders; and we prove new variants of the classical Parke-Taylor identities., Comment: 44 pages, 14 figures. v2: minor changes

A 2014 report on global human capital trends found that many employers across the United States had difficulties maintaining a skilled and committed workforce. Yet unemployment among young people ages 16 to 29 years was well above the overall unemployment rate. In 2016, 8.2 percent of this group were unemployed, while the national average was 4.9 percent. However, among Black young adults it was even higher: 13.8 percent. Indeed, unemployment among young Black men was more than triple the national average. To help address racial inequity and increase employment among young people, workforce development systems and other systems that serve young adults would need to change. Given this context, in 2016, the Annie E. Casey Foundation launched Generation Work™ to connect more young adults--especially young people of color from families with low incomes--with meaningful employment by changing the way public and private systems prepare them and support their search for jobs. As part of the initiative, it asked partnerships made up of key workforce organizations in five cities to change how their workforce systems operated by better aligning education, training, and support services across organizations, and weaving into their systems best practices related to employers and youth. This report examines the first five years of the initiative as implemented by five partnerships.

The amplituhedron is a mathematical object which was introduced to provide a geometric origin of scattering amplitudes in $\mathcal{N}=4$ super Yang Mills theory. It generalizes \emph{cyclic polytopes} and the \emph{positive Grassmannian}, and has a very rich combinatorics with connections to cluster algebras. In this article we provide a series of results about tiles and tilings of the $m=4$ amplituhedron. Firstly, we provide a full characterization of facets of BCFW tiles in terms of cluster variables for $\mbox{Gr}_{4,n}$. Secondly, we exhibit a tiling of the $m=4$ amplituhedron which involves a tile which does not come from the BCFW recurrence -- the \emph{spurion} tile, which also satisfies all cluster properties. Finally, strengthening the connection with cluster algebras, we show that each standard BCFW tile is the positive part of a cluster variety, which allows us to compute the canonical form of each such tile explicitly in terms of cluster variables for $\mbox{Gr}_{4,n}$. This paper is a companion to our previous paper ``Cluster algebras and tilings for the $m=4$ amplituhedron''., Comment: 44 pages, 20 figures. arXiv admin note: text overlap with arXiv:2310.17727

The Amplituhedron is a subspace of the Grassmannian that was recently defined by Arkani-Hamed and Trnka in their study of scattering amplitudes in planar $\mathcal{N}=4$ super Yang Mills theory (arXiv:1312.2007), and was the subject of many papers in the last decade. In this work we define a tropical analog of the amplituhedron, and develop techniques to address it. We prove that many of the key properties of the amplituhedron hold also in this simpler, piecewise linear, model., Comment: 44 pages, 2 figures

The papers [5, 3, 6, 19, 20] initiated the study of open $r$-spin and open FJRW intersection theories, and related them to integrable hierarchies and mirror symmetry. This paper uses a new technique, the point insertion technique, developed in the prequel [36], to define new open r-spin and open FJRW intersection theories. These new constructions provide potential candidates for theories whose existence was conjectured before: $\bullet$ K. Hori [23] predicted the existence of open $r$-spin theory with $\lfloor\frac{r}{2}\rfloor$ types of boundary states. The one constructed in [5, 3] has only one type of boundary state. In this work we describe $\lfloor\frac{r}{2}\rfloor$ open $r$-spin theories, labelled by $\mathfrak{h}\in\{0,\ldots,\lfloor\frac{r}{2}\rfloor-1\},$ where the $\mathfrak{h}$-th one has $\mathfrak{h}+1$ boundary states. We prove that the $\mathfrak{h}=0$ theory is equivalent to the [5, 3] construction, and calculate all intersection numbers for all these theories. $\bullet$ In [1] K. Aleshkin and C.C.M. Liu conjectured the existence of a quintic Fermat FJRW theory. We construct such an FJRW theory, and provide evidence that this is the conjectured theory. We also explain how the point insertion technique can be used for constructing other open enumerative theories, satisfying the same universal recursions.

A central component of rational behavior is logical inference: the process of determining which conclusions follow from a set of premises. Psychologists have documented several ways in which humans' inferences deviate from the rules of logic. Do language models, which are trained on text generated by humans, replicate such human biases, or are they able to overcome them? Focusing on the case of syllogisms -- inferences from two simple premises -- we show that, within the PaLM2 family of transformer language models, larger models are more logical than smaller ones, and also more logical than humans. At the same time, even the largest models make systematic errors, some of which mirror human reasoning biases: they show sensitivity to the (irrelevant) ordering of the variables in the syllogism, and draw confident but incorrect inferences from particular syllogisms (syllogistic fallacies). Overall, we find that language models often mimic the human biases included in their training data, but are able to overcome them in some cases., Comment: NAACL 2024

The amplituhedron $A_{n,k,m}(Z)$ is the image of the positive Grassmannian $Gr_{k,n}^{\geq 0}$ under the map ${Z}: Gr_{k,n}^{\geq 0} \to Gr_{k,k+m}$ induced by a positive linear map $Z:\mathbb{R}^n \to \mathbb{R}^{k+m}$. Motivated by a question of Hodges, Arkani-Hamed and Trnka introduced the amplituhedron as a geometric object whose tilings conjecturally encode the BCFW recursion for computing scattering amplitudes. More specifically, the expectation was that one can compute scattering amplitudes in ${N}=4$ SYM by tiling the $m=4$ amplituhedron $A_{n,k,4}(Z)$ - that is, decomposing the amplituhedron into `tiles' (closures of images of $4k$-dimensional cells of $Gr_{k,n}^{\geq 0}$ on which ${Z}$ is injective) - and summing the `volumes' of the tiles. In this article we prove two major conjectures about the $m=4$ amplituhedron: $i)$ the BCFW tiling conjecture, which says that any way of iterating the BCFW recurrence gives rise to a tiling of the amplituhedron $A_{n,k,4}(Z)$; $ii)$ the cluster adjacency conjecture for BCFW tiles, which says that facets of tiles are cut out by collections of compatible cluster variables for $Gr_{4,n}$. Moreover, we show that each BCFW tile is the subset of $Gr_{k, k+4}$ where certain cluster variables have particular signs. Along the way, we construct many explicit seeds for $Gr_{4,n}$ comprised of high-degree cluster variables, which may be of independent interest in the study of cluster algebras., Comment: v2: minor edits

The papers [3, 1, 4, 10] constructed an intersection theory on the moduli space of $r$-spin disks, and proved it satisfies mirror symmetry and relations with integrable hierarchies. That theory considered only disks with a single type of boundary state. In this work, we initiate the study of more general $r$-spin surfaces. We define the notion of graded $r$-spin surfaces with multiple internal and boundary states, and their moduli spaces. In $g=0,$ the disk case, we also define the associated open Witten bundle, and prove that the Witten bundle is canonically oriented relative to the moduli space. Moreover, we describe a method for gluing several moduli spaces along certain boundaries, show that gluing lifts to the Witten bundle and relative cotangent line bundles, and that the result is again canonically relatively oriented. In the sequel [20], we construct, based on the work of this paper, a family of $\lfloor r/2\rfloor$ intersection theories indexed by $\mathfrak{h}\in\{0,\ldots,\lfloor r/2\rfloor-1\},$ where the $\mathfrak{h}$-th one has $\mathfrak{h}+1$ possible boundary states, and calculate their intersection numbers. The $\mathfrak{h}=0$ theory is equivalent to the one constructed in [3, 1].

Doped heavy metal-free III-V semiconductor nanocrystal quantum dots are of great interest both from the fundamental aspects of doping in highly confined structures, and from the applicative side of utilizing such building blocks in the fabrication of p-n homojunction devices. InAs nanocrystals, that are of particular relevance for short wave IR detection and emission applications, manifest heavy n-type character poising a challenge for their transition to p-type behavior. We present p-type doping of InAs nanocrystals with Zn-enabling control over the charge carrier type in InAs QDs field effect transistors. The post-synthesis doping reaction mechanism is studied for Zn precursors with varying reactivity. Successful p-type doping was achieved by the more reactive precursor, diethylzinc. Substitutional doping by Zn2+ replacing In3+ is established by X-ray absorption spectroscopy analysis. Furthermore, enhanced near IR photoluminescence is observed due to surface passivation by Zn as indicated from elemental mapping utilizing high resolution electron microscopy corroborated by X-ray photoelectron spectroscopy study. The demonstrated ability to control the carrier type, along with the improved emission characteristics, paves the way towards fabrication of optoelectronic devices active in the short wave IR region utilizing heavy-metal free nanocrystal building blocks., Comment: 36 pages, 6 figures. SI- 17 pages, 10 figures

We study high dimensional expansion beyond simplicial complexes (posets) and focus on $q$-complexes which are complexes whose basic building blocks are linear spaces. We show that the complete $q$-complex (consists of all subspaces of a given linear space) may have non-trivial homology groups and therefore some techniques for simplicial complexes fail. We develop new techniques to work bypass this. In particular: (i) We describe a new construction of cones and use it to determine when the homology of the complete $q$-complex is trivial. We use this construction to prove the "projective support dimension conjecture" conjectured by Mnukhin and Siemons. (ii) We define topological high dimensional expansion for posets, and show that the complete $q$-complex has linear (in the number of lines) coboundary expansion. (iii) We define the $q$-Linial-Meshulam model of random $q$-complexes and prove a sharp threshold for the connectivity of random $q$-complexes., Comment: 27 pages, comments are welcome!

This brief highlights lessons from the City Colleges of Chicago (CCC) Centers of Excellence model, which has redesigned each of the system's seven campuses as a "college-to-career center" and consolidated academic programs in high-demand industries at particular campuses. The model incorporates features of sectoral training programs, which prepare people for quality jobs in specific industries and occupational clusters--sectors of the labor market--where there is strong local demand and the opportunity for career advancement. Evidence from studies of similar sector-focused training models outside of the community college setting suggests that the Centers of Excellence model could be a promising one for other large community colleges systems to emulate, and that it could even have important lessons for smaller systems and stand-alone community colleges. In a system in which each college focuses on a single sector or cluster of related sectors, colleges can be more responsive to employers, become the central point of contact for employers in a given sector, and, in the process, help ensure that students will obtain the skills and education that employers need and that will lead to good jobs with opportunities for career growth. This brief presents lessons from semistructured interviews conducted in 2021 with CCC faculty and staff members, employers, and community partners at two of the seven colleges in the system: Harry S Truman College and Richard J. Daley College. CCC administrators recommended that researchers focus on these two colleges because they represent two variations of the model: one Center of Excellence focus that is capital-intensive and one that is not. These qualitative research data were analyzed systematically to address the following primary research questions: (1) In what ways does the Centers of Excellence model help colleges deepen their knowledge of an industry and its needs??How, if at all, does the model facilitate connections with employers, industry associations, and public workforce agencies or help develop work-based learning opportunities for students; (2) How is the Centers of Excellence model put into operation in colleges??In what ways does it inform the development of career trajectories, services, and other forms of support provided to students in specific industries? and (3) What are the strengths of the Centers of Excellence model and what aspects could be further strengthened? What lessons does the Centers of Excellence model offer for other colleges and systems?

In this work, we present Conditional Adversarial Latent Models (CALM), an approach for generating diverse and directable behaviors for user-controlled interactive virtual characters. Using imitation learning, CALM learns a representation of movement that captures the complexity and diversity of human motion, and enables direct control over character movements. The approach jointly learns a control policy and a motion encoder that reconstructs key characteristics of a given motion without merely replicating it. The results show that CALM learns a semantic motion representation, enabling control over the generated motions and style-conditioning for higher-level task training. Once trained, the character can be controlled using intuitive interfaces, akin to those found in video games., Comment: Accepted to SIGGRAPH 2023

This brief highlights key findings from the implementation of the TechHire and Strengthening Working Families Initiative (SWFI) programs and offers considerations for practitioners. The programs provided training for high-tech jobs as well as support services to people with barriers to training and employment. Presented especially for a practitioner audience and derived from a rigorous evaluation, including visits to TechHire and SWFI programs, interviews with staff members, and a review of program participation data, the brief describes staff members' experiences of facilitators and challenges related to recruiting for these types of training programs and delivering a range of services, including technical skills training, case management, support services, child care assistance, and job placement and other post-training services. The brief offers considerations for practitioners involved in planning or implementing similar programs.

The stable gallium isotopes, $^{69,71}$Ga, are mostly produced by the weak slow ($s$) process in massive stars. We report here on measurements of astrophysically-relevant neutron capture cross sections of the $^{69,71}$Ga$(n,\gamma)$ reactions. The experiments were performed by the activation technique using a high-intensity ($3-5\times10^{10}$ n/s), quasi-Maxwellian neutron beam that closely mimics conditions of stellar $s$-process nucleosynthesis at $kT \approx$ 40 keV. The neutron field was produced by a mA proton beam at $E_p=1925$ keV (beam power of 2-3 kW) from the Soreq Applied Research Accelerator Facility (SARAF), bombarding the Liquid-Lithium Target (LiLiT). A 473 mg sample of Ga$_2$O$_3$ of natural isotopic composition was activated in the LiLiT neutron field and the activities of $^{70,72}$Ga were measured by decay counting via $\gamma$-spectrometry with a high-purity germanium detector. The Maxwellian-averaged cross sections at $kT$ = 30 keV of $^{69}$Ga and $^{71}$Ga determined in this work are 136(8) mb and 105(7) mb, respectively, in good agreement with previous experimental values. Astrophysical implications of the measurements are discussed.

In our previous two papers, we constructed an $r$-spin theory in genus zero for Riemann surfaces with boundary and fully determined the corresponding intersection numbers, providing an analogue of Witten's $r$-spin conjecture in genus zero in the open setting. In particular, we proved that the generating series of open $r$-spin intersection numbers is determined by the genus-zero part of a special solution of a certain extension of the Gelfand-Dickey hierarchy, and we conjectured that the whole solution controls the open $r$-spin intersection numbers in all genera, which do not yet have a geometric definition. In this paper, we provide geometric and algebraic evidence for the correctness of this conjecture., Comment: The content of this paper was included in the content of the first versions (v1-v4) of arXiv:1809.02536, before we splitted it. It includes a discussion of the structure of open r-spin intersection numbers in higher genera. 13 pages

Recent work in large language modeling (LLMs) has used fine-tuning to align outputs with the preferences of a prototypical user. This work assumes that human preferences are static and homogeneous across individuals, so that aligning to a a single "generic" user will confer more general alignment. Here, we embrace the heterogeneity of human preferences to consider a different challenge: how might a machine help people with diverse views find agreement? We fine-tune a 70 billion parameter LLM to generate statements that maximize the expected approval for a group of people with potentially diverse opinions. Human participants provide written opinions on thousands of questions touching on moral and political issues (e.g., "should we raise taxes on the rich?"), and rate the LLM's generated candidate consensus statements for agreement and quality. A reward model is then trained to predict individual preferences, enabling it to quantify and rank consensus statements in terms of their appeal to the overall group, defined according to different aggregation (social welfare) functions. The model produces consensus statements that are preferred by human users over those from prompted LLMs (>70%) and significantly outperforms a tight fine-tuned baseline that lacks the final ranking step. Further, our best model's consensus statements are preferred over the best human-generated opinions (>65%). We find that when we silently constructed consensus statements from only a subset of group members, those who were excluded were more likely to dissent, revealing the sensitivity of the consensus to individual contributions. These results highlight the potential to use LLMs to help groups of humans align their values with one another.

A researcher is interested in what sample size is needed to get the required significance of the same test, assuming exactly the same situation that was in the study with the non-significant result. We propose a simple solution to the problem., Comment: 14 pages including Appendix

As communication protocols evolve, datacenter network utilization increases. As a result, congestion is more frequent, causing higher latency and packet loss. Combined with the increasing complexity of workloads, manual design of congestion control (CC) algorithms becomes extremely difficult. This calls for the development of AI approaches to replace the human effort. Unfortunately, it is currently not possible to deploy AI models on network devices due to their limited computational capabilities. Here, we offer a solution to this problem by building a computationally-light solution based on a recent reinforcement learning CC algorithm [arXiv:2207.02295]. We reduce the inference time of RL-CC by x500 by distilling its complex neural network into decision trees. This transformation enables real-time inference within the $\mu$-sec decision-time requirement, with a negligible effect on quality. We deploy the transformed policy on NVIDIA NICs in a live cluster. Compared to popular CC algorithms used in production, RL-CC is the only method that performs well on all benchmarks tested over a large range of number of flows. It balances multiple metrics simultaneously: bandwidth, latency, and packet drops. These results suggest that data-driven methods for CC are feasible, challenging the prior belief that handcrafted heuristics are necessary to achieve optimal performance.

In this work, we aim to tackle the problem of autonomous grading, where a dozer is required to flatten an uneven area. In addition, we explore methods for bridging the gap between a simulated environment and real scenarios. We design both a realistic physical simulation and a scaled real prototype environment mimicking the real dozer dynamics and sensory information. We establish heuristics and learning strategies in order to solve the problem. Through extensive experimentation, we show that although heuristics are capable of tackling the problem in a clean and noise-free simulated environment, they fail catastrophically when facing real world scenarios. As the heuristics are capable of successfully solving the task in the simulated environment, we show they can be leveraged to guide a learning agent which can generalize and solve the task both in simulation and in a scaled prototype environment., Comment: 7 pages, Accepted to IEEE-IROS2022

Although approximately 50% of medical school graduates today are women, female physicians tend to be underrepresented in senior positions, make less money than their male counterparts and receive fewer promotions. There is a growing body of literature demonstrating gender bias in various forms of evaluation in medicine, but this work was mainly conducted by looking for specific words using fixed dictionaries such as LIWC and focused on recommendation letters. We use a dataset of written and quantitative assessments of medical student performance on individual shifts of work, collected across multiple institutions, to investigate the extent to which gender bias exists in a day-to-day context for medical students. We investigate differences in the narrative comments given to male and female students by both male or female faculty assessors, using a fine-tuned BERT model. This allows us to examine whether groups are written about in systematically different ways, without relying on hand-crafted wordlists or topic models. We compare these results to results from the traditional LIWC method and find that, although we find no evidence of group-level gender bias in this dataset, terms related to family and children are used more in feedback given to women., Comment: GeBNLP @ NAACL 2022